Tissue‐Specific Extracellular Vesicles Enriched From Circulation: Exploring the Liquid Biopsy Perspective

ABSTRACT

Extracellular vesicles (EVs) released from tissues can be found in human biofluids. EVs reflect the phenotypic state of their cell of origin, carrying informative molecular biomarkers within and between tissues, and represent a promising target for liquid biopsy. However, the heterogeneity of EVs in their surface composition, luminal content, biogenesis and cellular origin poses a challenge for selective enrichment and validation of origin‐specific EVs from the complex pool of circulating EVs. Another obstacle for translating EVs into liquid biopsy applications is the wide variety of separation and characterization methods, many of which lack standardization and reproducibility. In this review, we summarize current knowledge on tissue‐specific EVs, highlighting their potential as indicators of tissue health and disease, as well as the existing challenges and limitations. From the existing literature, we identify a compelling need for better validation and reproducibility studies to support the development of tissue‐specific EV applications. Identifying reliable tissue‐enriched biomarkers, in particular, will be required to enable further insights into the physiology and pathology of EV source tissues. We also propose some considerations to address future guidelines on the topic. Together, these approaches will help to establish EV liquid biopsy applications as a keystone of translational medicine.

1. Extracellular Vesicles

Extracellular vesicles (EVs) are small membrane‐delimited particles released by all known cells (Figure 1). They are present in various body fluids, such as blood, saliva, urine and breastmilk (Arraud et al. 2014; Kumar et al. 2024). EVs play key roles as mediators of cell‐to‐cell communication and signalling under both normal physiological and pathological conditions (Battistelli and Falcieri 2020; Pegtel and Gould 2019). The extensive crosstalk established across tissues makes EV biology critical for human health and disease (Pegtel and Gould 2019), and highlights their role in the adaptive capacity of cells in response to external stimuli (Bollati et al. 2023).

FIGURE 1.

Tissue‐specific EVs (TS‐EVs) in blood. Human blood is enriched in EVs derived from diverse tissues, facilitating extensive crosstalk across the body. TS‐EVs potentially reflect the phenotypic state of their releasing cells, under physiological or pathological conditions. They can serve as a promising tool for liquid biopsy applications by offering insight into hard‐accessible organs. Thanks to their heterogeneous molecular cargo, they also represent a source of tissue and cell‐specific biomarkers. Green colour represents EVs that have been targeted for capture from circulation in studies conducted to date, while blue colour denotes EVs that have been studied in vitro/in vivo and are likely present in circulation. TS‐EVs have been classified in the illustration for the purpose of the review, some organs may actually release multiple EVs subpopulations based on the multiple tissues they are composed of (Image created using Biorender.com).

EVs assume a fundamental role in several biological and pathological mechanisms like immune response and infections (Buzas 2023), metabolic and cardiovascular disease (Kalluri and LeBleu 2020), chronic inflammatory diseases (Harrell et al. 2019), neurodegeneration (Picca et al. 2022), tumour growth, metastasis and drug resistance (Bebelman et al. 2018; Kalluri 2016). Recent findings shed light on EVs involvement in mediating toxicity and how the exposure to toxicants (e.g., heavy metals, pesticides, ionizing radiations, cigarette smoke) can induce the alteration of their content (Rokad et al. 2019; Wu et al. 2023). The heterogeneous cargo content—including proteins, nucleic acids, lipids and small molecules—makes them a dynamic extension of their cells of origin (Kalra et al. 2016). Thus, as essential and ubiquitous cellular derivatives, EVs are able to shape the extracellular environment by transmitting signals and shuttling molecules to the nearby surroundings or at a distance (Pegtel and Gould 2019). Their ability to provide a snapshot of the source tissue, combined with their ease of accessibility from human biofluids, make EVs uniquely informative minimally‐invasive biomarkers (Wang et al. 2020). EVs can be classified according to their size, mechanisms of release and origins (Théry et al. 2018). Despite the specific nomenclature based on distinct features, the Minimal Information for Studies of Extracellular Vesicles (MISEV) guidelines recommend using the general term “EVs” to describe the entire category (Théry et al. 2018; Witwer and Théry 2019; Welsh et al. 2024). For all the reasons described above, the interest in EV enrichment and characterization has surged over the past decade in the scientific community, especially when referring to EVs circulating in biofluids (Gardiner et al. 2016; Raposo and Stoorvogel 2013).

2. Solid Tissue‐Derived EVs

Isolation of EVs directly from solid tissues captures primarily EV subpopulations residing in the interstitial space surrounding tissue cells and contains the molecular signals released by cells within the tissue's local communication network (Zhang et al. 2023; Crescitelli et al. 2021). These interstitial EVs preserve the complexity of the native microenvironment, capturing signals exchanged locally among the cells that form the tissue architecture (Lee et al. 2024; Li et al. 2021), as well as signals shaped by extracellular matrix interactions (Qin et al. 2021) and microenvironmental influences (Zhi et al. 2022). In contrast, tissue‐derived EVs in circulation represent the subset of vesicles that exit the tissue and are secreted into biofluids with properties that enable their release, stability and travel in the systemic compartment (Iannotta et al. 2024). These circulating EVs reflect signals intended for distal communication, carrying biologically meaningful cargo capable of influencing targets beyond their tissue of origin, or monocyte/macrophage uptake, mediating processes such as immune surveillance, inflammatory priming and tissue remodelling (Dietz et al. 2023; Czernek et al. 2015; Shimizu et al. 2021).

EVs released from a given tissue exhibit functional diversity, with distinct roles in the local microenvironment and in the systemic circulation (Li et al. 2021) and have potential different applications. Local tissue EVs can capture differences in adjacent normal and pathological cells, enabling contrasts that are often diluted once EVs enter the mixed circulating pool (Zhang et al. 2023). In contrast, the pool of circulating EVs integrate contributions from multiple cells and tissue sources, making them valuable informants of distal intercellular communication and useful as liquid biopsy tools, although they are less representative of tissue‐specific microenvironments compared with local interstitial EVs (Alberro et al. 2021; Zhang et al. 2023; Li et al. 2021).

Various methods have been reported for isolating EVs from tissues in both animal and human models, including the brain (Huang et al. 2020; Vella et al. 2017), liver (Matejovič et al. 2021), muscle (Leroyer et al. 2009), skeletal muscle and heart (Matejovič et al. 2021; Loyer et al. 2018), lymph nodes (Bodnár et al. 2025), thymus and spleen (Wang et al. 2008) and adipose tissue (Deng et al. 2009). Some of these methods have been applied to study neurological disorders (Gallart‐Palau et al. 2016; Hurwitz et al. 2019; Polanco et al. 2016) and tumour microenvironments (Hurwitz et al. 2019), as EVs from interstitial tissue space can modulate tumour microenvironment in cancer cell seeding and metastatic growth (Hoshino et al. 2015; Costa‐Silva et al. 2015). However, isolation and separation of solid tissue‐derived EVs are beyond the scope of this review. The present review mainly focuses on EVs enriched from circulation as powerful and promising avenue for liquid biopsy. The invasiveness of solid tissue sampling poses risks and discomfort to patients, is not always feasible, and despite the informative value of solid tissue‐derived EVs, its use is not an ideal approach for translational and precision medicine (Lehrich et al. 2024; Ilié and Hofman 2016).

3. EVs in Circulation as Targets of Liquid Biopsy

3.1. Liquid Biopsy

Liquid biopsy refers to the collection and analysis of body fluids as a minimally invasive method for early disease detection, monitoring disease progression, assessing therapeutic response, identifying genetic mutations and detecting residual disease, thereby reducing the need for solid tissue sampling (Poulet et al. 2019). Originally based on cell‐free DNA (cfDNA) and circulating tumour cells, the concept of liquid biopsy emerged a decade ago as a revolutionary technology for cancer molecular profiling and precision medicine treatment approaches. cfDNA‐based liquid biopsies reveal genetic and epigenetic features (Luo et al. 2021), whereas RNA and proteins can provide complementary or additional layers of information with higher tissue and cell specificity.

3.2. The Potential of Circulating EVs for Liquid Biopsy Applications

EVs have several characteristics that makes them well suited for use in liquid biopsy. They are relatively abundant in blood (serum and plasma), at an estimated 10¹⁰ particles/mL (Yu et al. 2022; Cai et al. 2015), and are mainly represented by small EVs (Buzas 2023). Carrying biological information from the parental cells, the lipid bilayer of EVs shields luminal cargo from degradation and enable crossing of biological barriers (Yu et al. 2022). Given the pleiotropic functions and the assorted cargoes, EVs‐based liquid biopsy can serve as a valuable tool to investigate cell status. EVs reflect the phenotypic state of the releasing cells (Buzas 2023; Kalluri 2016), and given the easy‐accessible sampling of the cargo, allow for an insight into the physio‐pathology of the source tissue (Kalluri and LeBleu 2020).

EVs encapsulate and disseminate circulating tumour DNA (Kalluri and LeBleu 2020), and EV nucleic acids allow more facile detection of common hotspot mutations in BRAF, KRAS and EGFR genes in patients with advanced cancers when compared with cfDNA (Möhrmann et al. 2018). EVs also contain cancer associated RNA and proteins, expanding the repertoire of available molecular markers (Möller and Lobb 2020; Onyiba et al. 2025). Analysis of cancer cell‐derived EVs may also provide a better overview of tumour heterogeneity than traditional tissue biopsies, which can be limited by the local sampling (Kalluri 2016).

3.3. Total EVs: A Systemic Overview

Blood and other biofluids are a complex mixture of EVs derived from multiple tissue. The analysis of total EVs released in a biofluids offers a comprehensive representation of biological processes throughout the body and provide a dynamic overview of its physiological/pathological changes (Alberro et al. 2021). However, the majority of EVs in circulation are from blood cells, and the proportion of EVs from different tissues is largely unknown and may vary under pathological conditions (Nieuwland and Siljander 2024). The analysis of total EVs in circulation is unable to discern the origin of the signals and the EV source, that is, tracing EVs back to individual parental cells to understand their specific phenotypic state.

In order to detect tissue molecular messages conveyed in circulation through EVs, and to develop a truly effective tool for biomarker discovery and liquid biopsy applications, tissue‐ and cell‐specific subpopulations of EVs must be captured. The analysis of individual subpopulation of EVs allows capturing informative molecules that may otherwise be overwhelmed by signals from irrelevant fractions (Van Der Pol et al. 2016). Biomarkers of interest may also be expressed by multiple tissues and an aggregated signal, as represented by total EVs, cannot provide information on cells of interest (Yu et al. 2022).

4. Tissue‐Specific EVs in Circulation

4.1. Disentangling Blood EV Complexity

The abundance of EVs in blood is highly variable, influenced by a number of biological and technical factors, posing a challenge for their analysis (Nieuwland and Siljander 2024; Zarovni et al. 2025). Physiologically, EVs abundance in blood may depend on the balance between secretion from each tissue and rapid uptake by recipients (Matsumoto et al. 2020). Technically, EV separation and quantification methods can substantially influence measured concentrations in biofluids. Moreover, co‐isolation of lipoproteins, platelets and residual blood cells can introduce contamination and biases (Nieuwland and Siljander 2024). Lipoproteins, for example, overlap with EV in size and density (Sódar et al. 2016). This is particularly important as lipoproteins level can be up to a million‐fold higher than EVs in blood, current separation methodologies cannot completely remove them, causing inaccuracies and variability in the estimates of EV abundance in blood (Nieuwland and Siljander 2024).

Experimental evidence showed that most EVs in blood originate from platelets or megakaryocytes (Van Der Pol et al. 2016; Berckmans et al. 2001; Flaumenhaft et al. 2009). Li et al. (2020) traced the source of tissue‐specific long RNA in circulating EVs in healthy subjects, applying a deconvolution algorithm to quantify the origin of plasma EVs (Li et al. 2019; Li et al. 2020). Their results suggest that most blood EVs originate from hematopoietic cells, while only a small percentage originate from other cell types. Of note, immune cells seem to release most of their EVs locally or in lymph nodes, which could explain their limited concentration in circulation (Lindenbergh and Stoorvogel 2018; Alberro et al. 2021). Beyond the hematopoietic component, their results suggested that adipose tissue is the predominant contributor, followed by muscle, lungs and liver (Li et al. 2020). To our knowledge this is the only study that computationally estimates the relative proportion of EVs from each source tissue, and the results need to be confirmed by experimental data. Beyond endogenous tissues, the complex mixture of blood EVs may also include EVs released by prokaryotic cells (Alberro et al. 2021).

4.2. The Role of Surface Markers

Cell surface receptors can be considered “tags” of the tissue of origin. Along with luminal cargo, EV surface composition reflects the molecular signature of the originating cell, deriving from its plasma membrane and endosomal compartments through mechanisms such as endosomal sorting (ESCRT, endosomal sorting complex required for transport) (Colombo et al. 2014). Combinations of cell surface receptors define “functional heterogeneity” of EVs, as different surface phenotypes may interact with specific cellular targets and convey different signals and information (Yáñez‐Mó et al. 2015). This EV “surfaceome” can also be used for EV identification, classification and separation (Hallal et al. 2022).

4.3. Tissue‐Specific EVs (TS‐EVs)

By leveraging the heterogeneity of the EV surfaceome, subpopulations of tissue‐specific EVs can be enriched from the EV pool in circulation. Preserving the information specific to the parental cell, EVs could potentially help to elucidate early pathological conditions across various tissues (Amin et al. 2024). Most of the knowledge about TS‐EVs, their functions and involvement in biological processes comes from in vitro studies (Royo et al. 2020). Recent literature reports multiple efforts to isolate and characterize TS‐EVs from human biofluids (Xie et al. 2025; Nogueras‐Ortiz et al. 2024; Bravo‐Miana et al. 2024). Indeed, according to the MISEV guidelines, in addition to the requirement of standardized separation methods, the study of EVs derived from specific sources (tissue/cells/biofluids) should include characterization of markers that indicate EV features and origins (Théry et al. 2018). There is a compelling need to identify and measure multiple biomarkers alongside commonly used ones, especially to investigate complex diseases (Califf 2018). Novel biomarkers from TS‐EVs might enhance predictive accuracy, adding information to those acquired by classical biomarkers of tissue functionality/damage.

5. Current Evidence on Circulating TS‐EVs Released From Major Tissues and Organs

5.1. Nervous System‐Derived EVs

Nervous System (NS) EVs are reported to be involved in cognitive functions by maintaining and improving synaptic plasticity and neurotransmission, as well as in neuronal development and neuroimmune communication (Qin 2020). NS EVs are released by neurons, astrocytes, oligodendrocytes, Schwann cells and microglia (Cano et al. 2023; Ghosh and Pearse 2023). Except for oligodendrocytes, which are restricted to the central (C)NS (their equivalents in the peripheral (P)NS are Schwann cells), all other neural cells are also found in the spinal cord and/or PNS in addition to the brain.

Prompted by findings that EVs could transfer misfolded proteins, such as amyloid‐beta, tau, alpha‐synuclein and TDP43, contributing to the propagation of neurodegenerative diseases (Witwer et al. 2013; Danzer et al. 2012), multiple efforts have sought to develop EV‐based diagnostic biomarkers and therapeutic strategies targeting neurological disorders (Yousif et al. 2022). The diagnostic potential of EVs for neurological disorders stems from their capacity to cross the blood–brain barrier into and out of the CNS (Chen et al. 2017). Thus, EVs from both the CNS and PNS can theoretically be isolated from biofluids such as blood, serving as a liquid biopsy for a tissue for which solid biopsy is rarely indicated (Picca et al. 2022; Mustapic et al. 2017). However, up to now, no markers or methods have been identified to distinguish EVs originating from the CNS or PNS.

Several studies have attempted to enrich neuron‐derived EVs by immunoaffinity procedures targeting neuronal surface antigens, especially neural cell‐adhesion molecule (NCAM) and L1 cell‐adhesion molecule (L1CAM), due to the high and relatively specific expression of these proteins in neural tissue (Fiandaca et al. 2015; Mustapic et al. 2017). L1CAM is the most common target for neuronal EV enrichment because of its high expression in neurons, but it is also detectable in hematopoietic and transformed epithelial cells (Nogueras‐Ortiz et al. 2024; Pulliam et al. 2019; Chen et al. 2018), as well as in kidney epithelial cells (Debiec et al. 1998), limiting its specificity. Some studies have reported difficulty reproducing the protocol. One study indicated that the primary forms of L1CAM protein in plasma and cerebrospinal fluids are truncated soluble isoforms, and that only a small fraction, if any, is present on the surface of EVs (Norman et al. 2021). This observation may not come as a surprise as L1CAM was described to be cleaved by several proteases, including presenilin/γ‐secretases and metalloproteases of the ADAM family (Maretzky et al. 2005). In contrast, Nogueras‐Ortiz et al. (2024) compared the use of L1CAM and other neuronal proteins including NCAM, Tau, NFL, GAP43, β‐III‐tubulin, VAMP2 and ENO2 to an isotype control and found L1CAM to be viable target for neuronal EV enrichment (Nogueras‐Ortiz et al. 2024). They demonstrated the presence of L1CAM on the surface of EVs via single EV analysis including fluorescence‐activated cell‐sorting (FACS) and immunogold electron microscopy. The presence of L1CAM on the surface of plasma EVs using immunogold electron microscopy has also been demonstrated by other independent groups (Shi et al. 2014; Yan et al. 2025; Jiang et al. 2020). Despite this, the debate about L1CAM remains open among researchers. Discrepancies observed between groups could potentially be explained by differences in protease inhibition strength and specificity between L1CAM EV isolation protocols. For instance, some protease inhibitor cocktails available commercially do not inhibit metal proteases.

Neuron‐derived EVs have also been immunocaptured using the sodium/potassium‐transporting ATPase subunit alpha‐3 (ATP1A3) (You et al. 2023), a combination of Growth Associated Protein 43 (GAP43) and Neuroligin 3 (NLGN3) (Eitan et al. 2023), and Neurexin‐3 (NRXN3) (Ter‐Ovanesyan et al. 2024). The specificity of these approaches was demonstrated by co‐isolating an extensive number of neuronal proteins (such as Tau, L1CAM, SYP, NRXN3), or RNAs (such as NRGN, ENO2 and NEFL), as well as by single EV analysis, immunogold microscopy, super‐resolution microscopy and other proteomic analyses. Data on efficiency and consistency were promising, but incomplete across the studies, and additional validation data may be needed. The low efficiency measured in those studies comparing the enriched and depleted fractions and by using a spike‐in recovery method, was suggestive of strong matrix effect. The consistency was only assessed between patients and controls within the study cohorts. Recently, we have performed a multi‐omic analysis of matched postmortem brain tissue, antemortem plasma total EVs and neuronal EV, targeting GAP43 and NLGN3, demonstrating better correlation between neuronal (GAP43+ and NLGN3+) EVs and brain tissue compared to total EVs and plasma (Kalia et al. 2025). However, we did not compare different methods (i.e., targets). These further highlights that systematic comparisons across approaches are still limited, underscoring the need for further validation studies to establish reliable and standardized methodologies. As it stands, while the optimal strategy for capturing neuronal EVs remains debated, existing methods have nonetheless provided valuable insights despite their limitations.

Several studies have applied neuronal EVs as a biomarker of neuronal functional state and to investigate the physiopathology of the brain (Mustapic et al. 2017; Cano et al. 2023). It has been reported that EVs can shuttle amyloid‐β and α‐synuclein, both implicated in Alzheimer's disease (AD) and Parkinson's disease (PD)(Winston et al. 2016; Shi et al. 2014; Eitan et al. 2023; Kapogiannis et al. 2019), suggesting utility as prognostic and diagnostic biomarkers (Delgado‐Peraza et al. 2021). Other luminal cargo molecules from neuronal EVs have also been shown to be possible sentinels of cognitive impairment in AD (Pulliam et al. 2019), preclinical AD (Fiandaca et al. 2015; Kapogiannis et al. 2019; Kapogiannis et al. 2015; Goetzl et al. 2015; Reho et al. 2025), cognitive impairment progressing to dementia (Winston et al. 2016) and disease severity in PD (Shi et al. 2014). Emerging studies have also suggested a role of neuronal EVs in the pathogenesis of lysosomal storage disorders (LSDs) manifesting neurological involvement and neurodegeneration (Tancini et al. 2019).

Efforts have been made to derive circulating EVs from neural cell types beyond neurons. Astrocyte EVs from blood of AD patients have been enriched by targeting protein markers such as glutamine aspartate transporter (GLAST; excitatory amino acid transporter 1 in humans/ EAAT1), glial fibrillary acidic protein (GFAP) and glutamine synthetase (GLuSyn) (Goetzl et al. 2016). EVs derived from microglia cells have been investigated in the context of fragility, targeting the protein transmembrane protein 119 (TMEM119) as specific marker (Visconte et al. 2023). Putative oligodendrocyte EVs have been studied in mouse serum targeting the 2′,3′‐cyclic nucleotide 3′‐phosphodiesterase (CNPase protein) (Zhang et al. 2024), and in human serum of Multiple Sclerosis (MS) patients targeting oligodendrocyte myelin glycoprotein (OMG) (Agliardi et al. 2023). Although applications of these other types of CNS EVs reported in the literature appear promising, the validation data on specificity, efficiency and consistency supporting their use are incomplete in most cases, leaving methodological rigor untested. Further investigations are needed to establish the reliability of the targets, especially for microglial and oligodendrocyte EVs, just as extensive validation studies have been conducted for neuronal EVs.

5.2. Liver‐Derived EVs

Hepatocytes constitute 80% of the total liver cell composition, representing the primary source of liver‐derived EVs. Other sources are endothelial cells and liver‐resident macrophages (Kupffer cells). Hepatic EVs contain enzymes and small non‐coding microRNAs involved in the metabolism of carbohydrates, lipids and xenobiotics (Newman et al. 2022). Liver EVs are released by hepatocytes under physiological state, but stress, damage, or pathological conditions can significantly affect their quantities and cargo composition (Wu et al. 2021; Xie et al. 2019). For example, hepatocyte EVs have been found to regulate vascular endothelial cell function when carrying the protein arginase‐1, altering the serum metabolites linked to oxidative stress (Royo et al. 2017). Holman et al. (2016) reported that stressed hepatocytes may release EVs as early immune responses during drug‐induced liver injury (Holman et al. 2016). They may contribute to local environmental responses to cellular or tissue injury and potentially play a role in repair or restoration (Maji et al. 2017).

Given the fundamental functions exerted by the organ, liver EVs have attracted considerable attention for their role in conditions such as non‐alcoholic fatty liver disease, alcoholic hepatitis, viral hepatitis, fibrosis and hepatobiliary tumours. A review by Muñoz‐Hernández et al. (2022) reports a compendium of surface/cargo markers identified in EVs from patients with liver diseases, despite the non‐specificity of the EVs isolated across the cited studies (Muñoz‐Hernández et al. 2022). Liu et al. (2023) identified MASP1 protein (mannan‐binding lectin serine protease 1) in EVs derived from hepatocytes as a potential biomarker of liver fibrosis in mice (Liu et al. 2023). Proteomic profiling of EVs secreted by hepatocytes identified several candidates with high potential as biomarkers for diagnosis, prognosis and treatment monitoring of liver diseases, such as ANXA2 (Annexin A2) for hepatocellular carcinoma (Urban et al. 2019; Conde‐Vancells et al. 2008). An in vivo model of hepatocellular carcinoma also identified proteins ASGR1 (Asialoglycoprotein receptor 1) and CYP2E1 (Cytochrome P450 Family 2 Subfamily E Member 1), which have been widely reported as specific to EVs of hepatocyte origin (Newman et al. 2022; Conde‐Vancells et al. 2008; Nakao et al. 2021; Povero et al. 2014). In agreement, a study targeting ASGR1 to specifically isolate liver‐derived EVs from patient's plasma, demonstrated that several miRNA biomarker candidates of liver disease measured in ASGR1⁺EVs, but not in total plasma RNA and total plasma EVs, were positively associated with disease severity and distinguished patients with non‐alcoholic fatty liver disease from its advanced form non‐alcoholic steatohepatitis (Newman et al. 2022). Other promising investigations on liver EV are ongoing (https://www.biospace.com/press‐releases/mursla‐bios‐evoliver‐surpasses‐current‐standards‐in‐liver‐cancer‐surveillance‐data‐presented‐at‐aasld‐liver‐meeting‐2024). Nevertheless, research into liver‐derived EVs is still a nascent field. Most EV studies involved data collected from total EVs in circulation, limiting insights into liver tissue EVs specifically, and the rare promising studies that reported methods to isolate liver‐specific EVs will need to be replicated on a larger scale.

5.3. Lung‐Derived EVs

EVs can be released naturally by lung cells or in response to stimuli such as inflammation or mechanical stress. Although TS‐EVs in the blood come from multiple sources and are mainly represented by platelets and hematopoietic cells, lung‐specific biological fluids, such as pleural fluid and bronchoalveolar lavage fluid (BALF), contain mostly EVs derived from lungs (McVey et al. 2019; Liu et al. 2022). Harvested from these sources, lung EVs are emerging as a component of multiple pulmonary pathologies (Liu et al. 2022), including lung cancer (Carreca et al. 2024), chronic obstructive pulmonary disease (Bartel et al. 2024), pulmonary arterial hypertension (Conti et al. 2023), asthma (Torregrosa Paredes et al. 2012), acute lung injury (Yuan et al. 2018) and acute respiratory distress syndrome (Zareba et al. 2021). Therefore, they appear to be promising biomarker candidates for lung health or disease, carrying miRNAs, other small non‐coding RNAs, and proteins, including cytokines (Mohan et al. 2020). Pulmonary tissue EVs are predominantly from epithelial cells, but a significant portion also come from endothelial cells, alveolar macrophages and fibroblasts (Fujita et al. 2015; Alipoor et al. 2016). Fine‐tuned regulation of EVs released by each type of lung cell is responsible for tissue homeostasis, balancing physio‐pathological scenarios. EVs derived from vascular endothelial cells and type II alveolar epithelial cells regulate the immune balance of alveolar macrophages and hold potential for therapeutic applications (Feng et al. 2021). Hyperoxia can stimulate the formation of lung epithelial‐derived EVs, prompting a pro‐inflammatory activation of macrophages; enriched in caspase‐3 cargo, these EVs can also mediate inflammatory lung response through oxidative stress‐induced apoptosis (Moon et al. 2015). Moreover, surface proteins from airway epithelial cell‐derived EVs, such as mucins, suggest a possible role in host immunity, while the presence of suppressor of cytokine signalling proteins (SOCS), carried by alveolar macrophage EVs, indicates a role in the regulation of cytokine responses and homeostasis of alveoli, controlling inflammatory signalling (Yamada 2021; Kesimer et al. 2009; Bourdonnay et al. 2015). There is evidence that EVs released from resident lung cells are actively involved in the homeostasis of the lungs and in host defence. The pivotal role of lung EVs in the pathogenesis of various bacterial and viral lung infections leads to regulation of the immune system and structural modification of lung cells (Hambo and Harb 2023).

Most research on lung EVs relies on markers obtained from total circulating EVs or in vitro cultures. A model of acute lung injury and acute respiratory distress syndrome, for example, identified the protein CD74 as a marker of EVs released by alveolar type II epithelial cells, targeting alveolar macrophages to exert pro‐inflammatory and anti‐fibrotic effects (Feng et al. 2021; Hu et al. 2022). An in vitro model of non‐small cell lung cancer (NSCLC), combined with patient tissue biopsies, reported CD91 to be a surface protein on EVs from both NSCLC cells in culture and human samples (Akbar et al. 2022). Lung EV research is an emerging field, and despite data from total EVs and in vitro models, EVs from pulmonary tissue have not been enriched from common and accessible biospecimen. Recently, the use of exhaled breath condensate has been reported as a means to enrich lung‐derived EVs, specifically those released by bronchiolar Clara cells and alveolar type II cells (Mitchell et al. 2024). However, from a liquid biopsy perspective, many efforts are still required to enrich lung‐derived EVs and develop them as markers of lung health or damage.

5.4. Adipose Tissue‐Derived EVs

Adipose tissue‐derived EVs can be released by mature adipocytes (white, brown and beige), adipose‐derived stromal/stem cells (ADSCs), preadipocytes, as well as by macrophages, endothelial cells and fibroblasts within adipose depots (Crewe 2022; Bond et al. 2022). Recent evidence shows that adipose tissue EVs contribute to local and systemic signalling that influence metabolic and cardiovascular cross‐talk (Michel 2023). In humans, circulating EVs enriched in adipocyte‐specific proteins such as perilipin A (PLIN1) and adiponectin, have been detected in plasma and correlated with insulin resistance and obesity (Connolly et al. 2018; Eguchi et al. 2016). In vitro/in vivo evidence show that adipose tissue‐derived EVs carry bioactive proteins and regulatory RNAs, such as fatty acid binding protein 4 (FABP4) and miR‐99b, modulating gene expression and metabolic signalling in distant organs, such as interfering with insulin signals in liver and muscle cells (Kranendonk et al. 2014), or linking adipose dysfunction to systemic metabolic disease (Thomou et al. 2017).

Adipocyte‐derived EVs have been enriched from circulation targeting FABP4 (Hubal et al. 2017), but unfortunately validation data demonstrating specificity and efficiency of this method were not provided. Mishra et al. (2023) used a panel of five proteins, CA3 (carbonic anhydrase 3), STEAP4 (six transmembrane epithelial antigen of prostate), FABP4 (fatty acid binding protein 4), GGT5 (gamma‐glutamyltransferase 5) and CAMKIIα (calcium/calmodulin‐dependent protein kinase type II subunit alpha) to target and capture adipocyte‐EVs from mice and human plasma. They validated the adipocytes‐EVs measuring adiponectin, adipokines and miRNAs related to obesity, comparing the enriched fractions of adipocyte‐EVs to the depleted ones. They showed that adipocyte EVs carry a strong pro‐inflammatory signal in diet‐induced obese mice compared to lean mice, and in the blood of nondiabetic obese individuals compared to non‐obese. They also show an increase in expression of obesity‐associated miRNAs and enrichment in obesity‐related metabolic pathways in the induced obese mice models compared to lean. However, the specificity of the capture was not demonstrated, as no isotype controls were included (Mishra et al. 2023). Collectively, these findings highlight the cellular diversity of adipose tissue‐derived EVs sources and their critical role as endocrine messengers linking adipose tissue physiology to whole‐body metabolic and cardiovascular regulation. However, validation data for the enrichment of adipose tissue–derived EVs from circulation remain limited, and further studies are needed to advance understanding in this area.

5.5. Heart‐Derived EVs

Heart‐derived EVs can be released by different cell types, such as cardiomyocytes, fibroblasts, endothelial cells, smooth muscle cells and cardiac progenitor cells, mediating intercellular communication and contributing to cardiac homeostasis and repair (Chistiakov et al. 2016). Most of the current knowledge on heart‐derived EVs comes from in vivo/in vitro studies. In vitro evidence on human induced pluripotent stem cell (iPSC)‐derived cardiomyocytes showed the release of EVs that are preferentially taken up by human endothelial cells, supporting their role in intercellular signalling and crosstalk in the myocardium (Zwi‐Dantsis et al. 2020). EVs collected from cultured human cardiac ventricular fibroblasts showed their role in the modulation of the calcium (Ca²⁺) cycling in human stem cell‐derived cardiomyocytes, a mechanism often altered in course of diseases, suggesting that EVs mediate of the fibroblast‐cardiomyocyte interaction (Wang et al. 2022). Intramyocardial injection of EVs derived from normal heart tissue have been shown to improve cardiac function after myocardial infarction, reduced scar size, inhibited cardiomyocyte apoptosis and enhanced angiogenesis in animal models of myocardial ischemia‐reperfusion. One mechanism involves delivery of mitochondrial protein ATP5a1 by EVs to preserve mitochondrial homeostasis and suppress cardiomyocyte ferroptosis in ischemia‐reperfusion injury (Liu et al. 2024).

A recently study, based on mass spectrometry and protein prediction approached, identified and targeted the proteins Popeye domain containing 2 (POPDC2) and Cholinergic receptor nicotinic epsilon subunit (CHRNE), as cardiomyocyte surface markers to facilitate the immunocapture of heart‐derived EVs from blood circulation in cellular, murine and human systems (Spanos et al. 2024). The study provides a quantitative, plasma‐based sample of the cardiomyocyte transcriptome, reflecting pathways related to human myocardial dysfunction, and capturing distinct pathways across health and disease conditions. The study was limited by the absence of validation data, which raises concerns about the specificity and efficiency of the approach. The promising data and applications for heart‐derived EVs are still limited and further studies are needed to better understand and advance methodologies.

5.6. Placenta‐Derived EVs

Placenta‐derived EVs are predominantly released by syncytiotrophoblasts directly into the maternal circulation through the uterine vein, through a process known as trophoblast deportation (Tong and Chamley 2015). They have recently garnered interest due to their role in feto–maternal communication, immune modulation and as biomarkers for pregnancy‐related disorders. Their potential in the early diagnosis, prevention and prognosis of pregnancy‐related conditions such as pre‐eclampsia, miscarriage, intrauterine growth restriction and gestational diabetes has been widely emphasized, although there are methodological challenges in their detection and purity (Buca et al. 2020). During pregnancies, EVs are released from the placenta, facilitating communication between the foetal–placental unit and maternal tissues (Tong and Chamley 2015). Placenta EVs enable a significant exchange of signals and molecules, with various cargo that may depend on the recipient targets (Buca et al. 2020; Ouyang et al. 2016). The release of placenta EVs increases over the course of gestation, and their bioactivity may change under disease conditions (Inagaki and Tachikawa 2022; Sarker et al. 2014). Feto–maternal communication includes cytokines and growth factors, modulating immune response and inflammatory stimulus during pregnancy (Germain et al. 2007; Fitzgerald et al. 2018), microRNA (Donker et al. 2012; Chaiwangyen et al. 2020), biologically active tRNA (Cooke et al. 2019) and trophoblastic phospholipids (Ouyang et al. 2016). Kupper and Huppertz (2022) described this intense crosstalk as the “endogenous exposome of the pregnant mother,” highlighting its crucial role in shaping the maternal system to meet pregnancy‐related demands (Kupper and Huppertz 2022). It also signifies that non‐invasive circulating biopsies may reflect placental health (Tannetta et al. 2017), warranting further investigation through robust and reproducible analytical approaches.

Placenta EVs express a set of placenta‐specific proteins that have been reported to mediate their internalization by maternal organs, including lungs, liver and brain (Inagaki and Tachikawa 2022; Kang et al. 2023). Most of these antigens have also been reported in the scientific literature as putative targets to enrich placenta EVs, including placental alkaline phosphatase (PLAP), Syncytin‐1 and Syncytin‐2, CD276 and human leukocyte antigen G (HLA‐G) (Vargas et al. 2014; Parveen et al. 2021; Kshirsagar et al. 2012). Some studies have found PLAP⁺ EVs increasing with gestational age in the serum of pregnant women (Sarker et al. 2014; Salomon et al. 2014), but immunocapture using PLAP fails to distinguish pregnant from non‐pregnant controls, which show nearly identical signal levels (Shinde et al. 2024). The authors attributed these results to non‐specificity of the antibody, due to placental alkaline phosphatase homology to other alkaline phosphatases (Sharma et al. 2014; Arai et al. 2003). The use of PLAP and HLA‐G together to enrich PLAP⁺/HLA‐G⁺ EVs resulted in a two‐fold increase in protein levels of HLA‐G in pregnant women compared with non‐pregnant controls (Shinde et al. 2024). Moreover, published protocols for placenta‐derived EV capture from maternal circulation, targeting both PLAP and HLA‐G by immunoaffinity, generally have not included isotype controls or measurement of tissue‐specific markers other than the capture antigen (Lai et al. 2018). In contrast, placenta EVs from primary cultures of human trophoblasts (Ouyang et al. 2016) and total EVs from the plasma of pregnant women have been immunophenotyped for placental origin by flow cytometry, using a panel of HLA‐G and syncytin‐1 antibodies (Ferrari et al. 2022). Based on the current literature and considering the potential of placenta EV applications to investigate feto–placental health, further efforts are needed to develop a more reliable and consistent method to enrich placental EVs.

5.7. Other TS‐EVs Found in Circulation: Bacterial‐Derived EVs

Both Gram‐negative and Gram‐positive bacteria release EVs into the extracellular environment, contributing to bacteria‐bacteria and host‐bacteria interaction (Díaz‐Garrido et al. 2021). Although strictly not a “tissue”, the unique properties of bacterial EVs position them as functionally relevant EVs in biofluids. Bacterial EVs shape the composition of microbial communities, with functions ranging from weakening or killing competitors to cooperating in adhesion and invasion of host tissue in the case of infections (Kim et al. 2015). Bacterial EVs also serve as a defence mechanism against harmful agents, such as antibiotics, removing toxic compounds from the bacterial cell (Wettstadt 2020). In host‐bacteria communication, bacterial EVs modulate host cell immune responses (strains that establish a symbiotic relationship) or contribute to immune evasion (infective species) (Kim et al. 2015), promoting non‐immunogenic, pro‐inflammatory, or cytotoxic responses, depending on the specific characteristics of the released EV (Ünal et al. 2011; Badia and Baldomà 2020). The cargo of bacterial EVs depends on the source bacteria, growth phase, environmental conditions and biogenesis process, suggesting a high selectivity process (Díaz‐Garrido et al. 2021; Toyofuku et al. 2023). Microbial EVs may be particularly relevant in determining the intestinal microbial environment in inflammatory bowel diseases (Shen et al. 2022); as messengers in the gut‐brain axis that may contribute to neurological disorders (Sun et al. 2023); and in the context of lung cell‐microbiome interactions during respiratory infection (Park et al. 2023).

The enrichment of bacterial EVs from host biofluids is particularly challenging due to the high heterogeneity of the bacterial community itself (Gurunathan and Kim 2023) and potential bacterial contamination, which may occur during blood collection and subsequent handling and processing. Several studies have phylogenetically characterized putative bacterial EVs in circulation through 16S rRNA sequencing of total EVs (Samra et al. 2021; Northrop‐Albrecht et al. 2022; Lee et al. 2021; Su et al. 2022; Kang et al. 2023). Bacterial EVs have also been specifically captured from human faeces and blood plasma (Xue et al. 2023; Fizanne et al. 2023; Tulkens et al. 2020). Some studies rely on the lipopolysaccharide (LPS) and lipoteichoic acid (LTA) to distinguish between Gram‐negative and Gram‐positive bacterial EVs, respectively (Ñahui Palomino et al. 2021). A study by Tulkens et al. (2020) investigated Gram‐negative EVs from plasma of patients with intestinal barrier disfunction, enriching the EV fraction by density gradient ultracentrifugation (specific density ∼1.141 to 1.186 g/mL) (Tulkens et al. 2020). They measured the activity of bacterial EV‐associated LPS related to the systemic immune activation of the hosts, showing that the levels of pathogen‐associated molecular patterns (PAMP), such as Tool like receptor 4 (TLR4), Zonulin protein and proinflammatory cytokines, were correlated with impaired barrier integrity in patients with inflammatory bowel disease (IBD), human immunodeficiency virus (HIV) and cancer therapy‐induced intestinal mucositis. It has also been reported that LPS⁺ bacterial EVs reach the bloodstream, accumulate in the liver and trigger inflammation through TLR4‐mediated signalling pathways (Jain et al. 2024). As an emerging field, knowledge and advancement may help elucidate new therapeutics and diagnostics based on bacterial EVs and enhance understanding of the interplay that defines the co‐existence of host and the bacterial communities.

5.8. Multiple Efforts Lacking Consistent Validations

So far, the TS‐EV research subfield has overall suffered from a lack of consistency in methodological approaches, particularly an absence of supporting validation to establish the reliability and specificity of the capture process from biofluids. For example, several studies that target surface markers to select TS‐EVs specifically, use the same surface proteins for the enrichment validation, without also validating the results with luminal TS‐EV markers (Visconte et al. 2023; Lai et al. 2018; Dutta et al. 2021). Moreover, most of the published literature does not report the use of isotype controls in the process of TS‐EV enrichment to assess specificity (Visconte et al. 2023; Agliardi et al. 2023). Lack of full methodological details to establish specificity, efficiency and consistency can hinder experimental reproducibility (Nieuwland and Siljander 2024). These limitations are an obstacle to the translation of tissue‐specific EVs into liquid biopsy applications. Additional considerations are discussed in Section 8.

6. Methods for Enrichment and Characterization of Tissue‐Specific EVs From Circulation

The enrichment and characterization of TS‐EVs from circulation are particularly challenging due to EV heterogeneity, the compositional complexity of biological fluids and the lack of standardized approaches. Most of the commonly used separation methods yield a fraction enriched with certain EV subpopulations, but these are often contaminated by non‐target EVs and other extracellular particles (EPs) (Jankovičová et al. 2020). One class of separation methods could be defined by reliance on physical properties of EVs. Whether using density‐based methods (such as density gradient centrifugation), size‐based methods (such as size exclusion chromatography or ultrafiltration), methods that depend on both size and density (ultracentrifugation), or charge‐based methods (ion exchange chromatography), the resulting EV‐enriched fraction will include other EPs that share physical characteristics with the EVs (Böing et al. 2014; Théry et al. 2006; Fortunato et al. 2022). Combination of several such isolation methodologies can increase EV purity, but with large reduction in yield as well, especially since several potential contaminations also interact with EVs (Corona proteins) (Tóth et al. 2021). Since separation strategies based solely on physical properties do not enable the enrichment of TS‐EVs, methodologies to capture distinctive surface phenotypes are needed (Mondal and Whiteside 2021). However, combining standard EV purification methods with techniques to enrich EV subpopulations of interest can yield a higher‐purity final product, reducing the contamination of lipoproteins or plasma proteins that can interfere with the characterization of the targeted TS‐EVs (Brennan et al. 2020).

The enrichment of TS‐EVs from circulation differs from the detection of TS‐EVs in circulation. The enrichment of TS‐EVs actively isolates and separates a subpopulation of interest from the pool of total EVs in circulation, and coupled with downstream molecular analysis, facilitates the profiling of the cargo. In contrast, the detection of TS‐EVs can measure their presence or specific markers directly within a biofluid, providing an overview of abundance and origin without physically isolating them or further characterization (Newman and Rowland 2025). However, the TS‐EVs markers detected are not always compatible with enrichment methods of TS‐EV subpopulations. Further considerations are in paragraph 6.2.

6.1. Immunoaffinity‐Based Enrichment of Circulating TS‐EVs

Immunoaffinity capture methods are based on the specific binding between selected surface antigens and their ligands, such as protein‐antibody interactions. Other affinity‐based approaches use DNA aptamers and peptides (Gaillard et al. 2020; Yi et al. 2021) Ligands can be immobilized on or conjugated to several solid supports, such as plates, magnetic beads, microfluidic devices, or polymeric materials. The high purity obtained, despite the relatively low recovery, makes the approach suitable for enriching EVs derived from specific sources (Gao et al. 2023; Doyle and Wang 2019). The most widely used immunoaffinity capture method uses biotinylated antibodies against EV surface proteins, covalently coupled to magnetic beads (Yousif et al. 2022; Nogueras‐Ortiz et al. 2024; Brett et al. 2017; Eitan et al. 2023).

6.2. The Specificity of Surface Proteins

Each EV expresses distinct and varied surface markers that reflect the releasing cell type. Those markers can be useful to trace back EV origins and selectively enrich the fraction of interest from complex biofluids (Kowal et al. 2016). Theoretically, any protein or cell membrane component that is exclusively or predominantly expressed on the surface of the TS‐EVs and lacks soluble counterparts in extracellular fluids can be used for immunoaffinity‐based EV capture (Yang et al. 2020). Practically, identifying suitable targets for EV immunocapture is challenging because, beyond the mere cell‐specificity of the markers, it is unclear which of those are enriched or depleted on the surface of the TS‐EV (Figure 2) and which of the available antibodies, if any, is the most efficient for the EV immunoaffinity application. Distribution of cell‐specific markers can vary in the total EV population and among the same EV subtype (Zarovni et al. 2025). EV surfaceosome topology can also be a determinant of the targetability of TS‐EV markers (Hallal et al. 2022). The properties of the EV membrane, including lipid orientation and distribution, can influence surface proteins aggregation and interactions, and consequently their exposure to solute and availability for binding antibodies. Furthermore, stoichiometry of EV membrane proteins is critical, since it can affect the number of available copies of marker to be targeted by an antibody. Small EVs are reported to be composed by more membrane than luminal proteins (Zendrini et al. 2022). Some evidence suggests that different classes of proteins can be differentially expressed on the surface of a single EV, with a number of copies spanning approximately from 2–3 to 200 for each EV (Zarovni et al. 2025). For these reasons, some cell‐specific proteins could be underrepresented in the EV subpopulation of interest, making them immunocapture extremely difficult and their detection biased by the analytical technique employed.

FIGURE 2.

Illustration of TS‐EV surface markers. EVs express surface markers reflecting the cell of origin. The figure shows the plasmatic membrane of a source cell, characterized by common membrane proteins (transmembrane, glycoprotein, channel, integral) and cell‐specific surface markers “A”, “B”, and “C”. In the example, EV subtypes released by the cell are enriched for the marker A (Marker A⁺ EV), for a combination of markers A and B (Marker A⁺/B⁺ EV), or markers B and C (Marker B⁺/C⁺ EV). As in the case of the subpopulation B⁺/C⁺ EV, compared to the source cell, the cell‐specific marker A is depleted on the surface of the EV, while marker C is expressed but under‐represented (only one copy of the protein). This could make the marker C not an ideal candidate as identification/classification/separation target (Kowal et al. 2016; Yang et al. 2020; Zarovni et al. 2025). (Image created using Biorender.com).

6.3. Immunoaffinity‐Based Enrichment of Circulating TS‐EVs: Evidence From Literature

Literature reports examples of TS‐EV enrichment from circulation through the immuno‐affinity capture methods (Table 1). Anti‐EpCAM‐conjugated magnetic beads have been used to target and isolate tumour‐derived EVs, due to EpCAM overexpression in different cancer cells (Königsberg et al. 2011). As covered above, neuronal EVs have been extensively investigated to identify prognostic and diagnostic biomarkers of neurological disorders, targeting L1CAM protein, or GAP43 and NLGN3 (Kapogiannis 2017; Gomes and Witwer 2022; Eitan et al. 2023; Yousif et al. 2022). Antibodies anti‐CD61 and anti‐CD41 have been used to isolate platelet EVs (Wang et al. 2016), while anti‐PSMA (anti‐prostate specific membrane antigen) have been considered for prostate EVs (Brett et al. 2017). Mishra et al. (2023) used a cocktail of five antibodies, based on a panel of mass‐spectrometry‐identified surface markers, to selectively enrich adipocyte EVs from mouse and human plasma: CA3, STEAP4, FABP4, GGT5 and CAMKIIα (Mishra et al. 2023). The success of the immunoaffinity capture approach for TS‐EVs largely depends on (i) precise identification of cell‐specific surface protein markers, (ii) availability of reliable antibodies and (iii) minimization of non‐specific interactions (Stam et al. 2021). A robust and appropriate capture method is foundational to EV‐based liquid biopsy research and applications. Currently, the absence of a standardized and reliable approach, and thus the inability to efficiently capture and enrich TS‐EVs, is a major barrier to replicability and clinical translation (Newman et al. 2022). Some studies rely on the detection and immunophenotyping of the TS‐EV subpopulations of interest in circulation, instead of using enrichment approaches (Ferrari et al. 2022; Newman and Rowland 2025).

TABLE 1.

Enrichment of TS‐EVs from body fluids by immunoaffinity capture. Evidence from current literature.

Source tissue/organ	Cell type	Surface marker	References
Nervous system	Neurons	L1CAM N1CAM GAP43 NLGN3 ATP1A3 NRXN3	(Mustapic et al. 2017) (Fiandaca et al. 2015) (Eitan et al. 2023) (Eitan et al. 2023) (You et al. 2023) (Ter‐Ovanesyan et al. 2024)
	Astrocytes	GLAST (EAAT1)	(Goetzl et al. 2016)
	Microglia cells	TMEM119	(Visconte et al. 2023)
	Oligodendrocytes	CNPase OMG PDGFRα	(Zhang et al. 2024) (Agliardi et al. 2023) (Kumar et al. 2023)
Liver	Hepatocytes	ASGR1	(Rodrigues et al. 2021)
Lung	Club cells, alveolar type 2 cells	CCSP SFTPC	(Mitchell et al. 2024)
Placenta	Syncytiotrophoblasts	PLAP HLA‐G	(Shinde et al. 2024) (Lai et al. 2018)
Adipose	White adipocytes	FABP4 CA3 STEAP4 GGT5 CAMKIIα	(Hubal et al. 2017) (Mishra et al. 2023)
Intestine	Intestinal epithelial cells	GPA33 CLRN3 GCNT3 PIGY REG4	(Nazarova et al. 2021)
Skeletal muscle	Myocytes	SGCA	(Guescini et al. 2015)
Heart	Cardiomyocytes	POPDC2 CHRNE	(Spanos et al. 2024)
Blood	Platelets	CD61 CD41	(Wang et al. 2016)
Prostate	Prostate epithelial cells	PMSA	(Brett et al. 2017)

6.4. Characterizing Circulating TS‐EVs by Flow‐Cytometry Analysis

Nanoparticle tracking analysis (NTA), dynamic light scattering (DLS) and transmission electron microscopy (TEM) can provide valuable information on EV general characteristics, including size, concentration and morphology (Omrani et al. 2024). However, these techniques fall short in addressing the multilevel heterogeneity of TS‐EVs. In contrast, flow cytometry (FC) can help identify and analyse EV subpopulations, including markers of their cellular origin (Welsh et al. 2023; Brealey et al. 2024). Given the small size and heterogeneity of EVs, traditional FC has faced challenges to accurately detect vesicles sized <500 nm. However, recent advancements in technology, such as the development of high‐resolution instruments and optimized protocols, have significantly improved the quality of the detection, defining the potential utility of FC for the characterization of TS‐EVs (Alberro et al. 2021; van der Vlist et al. 2012; Morales‐Kastresana and Jones 2017). FC allows the detection and quantification of surface markers specific to TS‐EV parental cells (Gul et al. 2022). The immunophenotyping approach by FC measures the presence of surface antigens using fluorescence‐labelled antibodies (Coumans et al. 2017). It enables the analysis of thousands of EVs in one sample, simultaneously determining multiple markers (Szatanek et al. 2017). Moreover, bead‐based detection flow‐cytometric assay has been developed to detect EV surface signatures with most standard FC devices (Wiklander et al. 2018). Additional published workflows provide methodologies suitable for fresh blood applications, avoiding analytical pre‐steps, such as imaging flow cytometry (IFC) (Woud et al. 2022) and a combination of polychromatic flow cytometry (PFC) with fluorescence‐activated cell‐sorting (FACS) (Marchisio et al. 2020). A recent review paper highlighted the ongoing efforts to standardize FC approaches for EV analysis and underscoring the progress made in enhancing the accuracy and reliability of the technology applied to EV research (Bettin et al. 2023). Many researchers are also focusing their efforts beyond FC technology on improving methodologies of label‐free imaging to characterize specific subpopulation of EVs (Leggio et al. 2023). Overall, we see the combination of single EV and bulk EV methodologies as powerful in identifying and validate TS‐EVs.

7. Characterizing EV Cargo Can Help Validate Tissue of Origin

High‐throughput omics technologies, including mass spectrometry‐based proteomics (Pathan et al. 2019) and lipidomics, as well as microarray‐ and next‐generation sequencing‐based transcriptomics, have been used to characterize EV cargo (Kim et al. 2015). The nature of the EV total cargo depends on the physiological or pathological state of the donor cell, as well as on stimuli that modulate vesicle biogenesis, production and release (Van Niel et al. 2018). To investigate the complexity of EV cargo, web‐based repositories such as Vesiclepedia (http://www.microvesicles.org/) and EVpedia (https://ngdc.cncb.ac.cn/databasecommons/) are curated compendia of EV‐related protein data (Kalra et al. 2016; Pathan et al. 2019; Kim et al. 2013; Kim et al. 2015; Chitti et al. 2024).

7.1. Common EV Markers Coexist With Specific Signals From the Parental Tissue

Proteomic characterization studies reveal that the majority of EV proteins are common to many cell types, including proteins involved in EV biogenesis (Hurwitz et al. 2016; Kugeratski et al. 2021), such as tetraspanins (i.e., CD9, CD81, CD63, CD82), signal transduction proteins (i.e., protein kinases, G proteins), chaperones (i.e., HSP70, HSP90) and intracellular trafficking proteins (i.e., RAB, GTPases, annexins). Only a small fraction of the EV proteome is cell‐type specific (Van Niel et al. 2018; Rädler et al. 2023). Similarly, published transcriptomic data suggests that EV RNAs reflect the cell of origin, but differ from cellular RNA contents in terms of RNA species and relative concentration of specific RNA sequences (O'Brien et al. 2020; Sork et al. 2018).

7.2. EV Total Cargo Signature That is Tissue‐Specific

A systematic assessment of small RNA profiling revealed that specific human cancer cells sort functionally important miRNAs into EVs, suggesting that these miRNAs might be part of molecular signalling pathways that drive tumorigenesis (Wang et al. 2023). From the EV RNA transcriptomic landscape emerged that EVs carry unannotated small RNAs associated with tissue‐specific phenotypes of advanced prostate cancer (Dogra et al. 2024). Similarly, a comparative proteomic analysis of macrophage‐derived tumour EVs identified a specific signature of 62 proteins (Cianciaruso et al. 2019). Microglia‐derived EV cargo is characterized by protein, mRNA and microRNA signatures that are specific for the EV subpopulation and depend on the microglial cell activation state (Santiago et al. 2023). Distinctive protein signatures associated with each EV organ of origin were identified by mass spectrometry proteomics of mouse tissues, additionally confirmed by detection in circulating blood EVs (Abdelmohsen et al. 2023).

8. Considerations on Tissue‐Specific EVs Nomenclature

Nomenclature of TS‐EVs has been a controversial point in the field. The MISEV guidelines suggest the use of the general term “EV” to include all the categories that are differentiated based on size, features and origin, but also extend the nomenclature to the “operational terms” that can be added as a prefix to the term “EV” (Théry et al. 2018; Welsh et al. 2024). However, there is little guidance or discussion in existing literature on properly naming TS‐EVs. Currently, authors refer to the tissue‐specific or ‐derived or‐enriched EVs isolated from a biofluid based on the tissue or cell of origin (i.e., neuronal EV), or on the marker employed for the enrichment (i.e., L1CAM⁺ EV). The two nomenclatures should not be considered synonymous with each other.

A surface marker can be expressed by different cells; thus, EVs purified by targeting a marker could partially derive from multiple types of source cells. This is the case for L1CAM, where the protein is highly expressed by neurons, but it is also detectable in other cells (i.e., hematopoietic cells, kidney cancer cells) (Pulliam et al. 2019). Likewise, PLAP is enriched in placenta but is also expressed in the gastrointestinal tract. TMEM119 is a shared marker between macrophages and microglia cells (Uhlén et al. 2015). Therefore, it may not be accurate to label an EV enriched by an anti‐L1CAM antibody as “neuron‐derived,” even though neurons may be the primary source of L1CAM⁺ EVs (Nogueras‐Ortiz et al. 2024; Shinde et al. 2024); or “placenta‐derived” using the anti‐PLAP antibody; or “microglia‐derived” using the anti‐TMEM119 antibody. EVs “enriched from” neuronal, placental or microglial origin appear more appropriate but could be insufficient. Moreover, EVs derived from a specific tissue do not have universally identical surface markers. Many molecules are integrated into the EV membrane bilayer or surround it as a coating protein corona (Hallal et al. 2022; Tóth et al. 2021). L1CAM, for instance, is not the only antigen expressed on EVs released from neurons. Purifying EVs targeting L1CAM results in a subpopulation of neuronal EVs (mostly) that display L1CAM, but other EVs derived from neurons may be negative for L1CAM, while carrying different antigens (e.g., NLGN3⁺, NRXN3⁺ or ATP1A3⁺) (You et al. 2023; Eitan et al. 2023; Ter‐Ovanesyan et al. 2024).

Altogether, we suggest that the most valid and effective way to define TS‐EVs is by the marker targeted for the enrichment of a particular subpopulation. Adopting a consensus about TS‐EV terminology could help resolve wide confusion in the TS‐EV research field, which arises from the use of different terms to describe the same concepts. This could also help improve reproducibility across studies.

9. General Guidelines for Validation of TS‐EVs

This section is intended to provide practical suggestions to approach the study of TS‐EVs (Figure 3). To date, TS‐EV separation methods are based on affinity capture, targeting surface antigens mostly by antibodies. The performance of these methods can be evaluated by measuring specificity, efficiency and consistency parameters (Box 1). It is important to highlight that we do not suggest that all strategies below must be performed for each assay in development; rather, they serve as a framework suited to the stage and purpose of development. Moreover, the following is not an analytical validation guideline, which should be developed for each assay according to regulatory guidelines.

FIGURE 3.

Key concepts to approach the study of TS‐EVs. NTA, nanoparticle tracking analysis; TEM, transmission electron microscopy; TS‐EVs, tissue‐specific extracellular vesicles. (Image created using Biorender.com).

Box 1: Highlights on method performance parameters

9.1. Specificity

Measurement of the amount of target captured by the specific agent in comparison to a control non‐specific agent, in the case of antibodies, isotype control (e.g., IgG) is the first step to verify specificity. This also confirms the source of the isolated EV subpopulation. In addition, the antibody quality can be validated using different methodologies (Weller 2018; Jarmoskaite et al. 2020). EVs are a package of macromolecules, and thus specificity can be measured by markers that should co‐isolate with the affinity agent target. The additional marker can be a second surface protein, a luminal protein, RNA, or lipid. The more co‐isolated markers and the more diverse modalities probed, the more substantial the evidence for TS‐EV specificity. In addition to measuring specific molecules, it is possible to perform an unbiased test using methods like RNAseq and proteomic. Results can be analysed against a set of transcripts or proteins known to be enriched in a specific tissue (Pei et al. 2019), or by gene ontology analysis. If possible, the TS‐EV cargo can be compared to its tissue of origin (Kalia et al. 2025; Delgado‐Peraza et al. 2021).

EVs are highly heterogeneous, and a single cell type can secrete a diverse EV population. Specificity and frequency of a marker for an EV subpopulation can be determined by single‐EV analysis. Single‐EV flow cytometry can be used to determine the percentage of EVs that contain a specific marker. However, FC suffers from low sensitivity, as they generally only detect EVs with several markers and larger than 70 nm. Super‐resolution microscopy and immunogold electron microscopy can detect single markers on the surface of EVs. However, not all antibodies are compatible with these methods, and quantifiability is limited.

9.2. Efficiency

The most common method to measure the efficiency of affinity isolation is to compare the starting biospecimen, the captured fraction and the depleted unbound fraction. However, this can be challenging for blood/plasma/serum due to the complex matrix effect, which can have a strong confounding impact on immunoassays. Therefore, these types of data should be interpreted carefully. Spike‐in recovery is also a common methodology for analysing isolation efficiency. EVs from cell lines, tissues, or any other source that represents the target cell type should be isolated and quantified. Then, a known amount of EV can be spiked into plasma samples, and recovery can be analysed.

9.3. Consistency

With plasma samples, there are two types of consistency: technical consistency and matrix interference due to biological and/or pre‐analytical variability. Technical variability can be addressed by running the same procedure multiple times and calculating the variability. However, with blood/plasma/serum samples, it is also important to verify specificity and efficiency across several individual plasma samples as large variability can occur. It is also recommended that these analyses be repeated for any new cohort to ensure the assay is compatible with the specific cohort's pre‐analytical conditions.

General EV characterization should also be considered to outline EV common features. Omics approaches (transcriptomics, proteomics, metabolomics, lipidomics) can also be pursued to analyse the EV cargo content and define a tissue‐specific profile, validating the origin of the subpopulation.

10. Conclusions: Limitations and Challenges

Tissue‐specific EVs as a liquid biopsy enabling tool represents a developing field within translational medicine, providing a cutting‐edge method to enhance the understanding of disease biology and advancing precision medicine paradigms. However, there are still many questions open around their potential. Exploring the features of each TS‐EV, in physiological or pathological conditions, may help clarify the activated pathways and the role exerted to the target tissues and the range of possible clinical applications. Currently, one major limitation to the use of TS‐EVs for liquid biopsy applications is the lack of standardized methodologies to enrich and characterize them from complex biofluids, although several promising approaches are emerging. Despite the current limitations of TS‐EVs as liquid biopsy markers, further research should be pursued.

Author Contributions

Biancamaria Pierri: conceptualization, writing – original draft, writing – review and editing. Erez Eitan: conceptualization, writing – review and editing. Kenneth W. Witwer: conceptualization, writing – review and editing. Diane B. Re: writing – review and editing. Andrea A. Baccarelli: writing – review and editing. Haotian Wu: conceptualization, supervision, writing – review and editing.

Funding

This work was supported by the National Institute of Environmental Health Sciences (NIEHS) under Grant number 1R35ES031688‐01A1, R01ES032242, R01ES029971.

Conflicts of Interest

Dr. Erez Eitan is employed at NeuroDex Inc. The other authors declare no conflicts of interest.

Pierri, B. , Eitan E., Witwer K. W., Re D. B., Baccarelli A. A., and Wu H.. 2026. “Tissue‐Specific Extracellular Vesicles Enriched From Circulation: Exploring the Liquid Biopsy Perspective.” Journal of Extracellular Biology 5, no. 2: e70106. 10.1002/jex2.70106

Data Availability Statement

Data sharing is not applicable to this article as no datasets were generated or analysed during the current study.