Exosomal bioactive molecules: molecular regulation, functional heterogeneity, and translational implications in disease diagnostics and therapeutics

Abstract

Exosomes are nanoscale extracellular vesicles that play a pivotal role in cell-to-cell communication by transporting a rich cargo of bioactive molecules, including proteins, lipids, and diverse nucleic acids. Recent discoveries have uncovered the complex molecular machinery behind exosome biogenesis and cargo loading, highlighting the roles of Endosomal Sorting Complex Required for Transport (ESCRT) complexes, lipid domains, RNA-binding proteins, and post-translational modifications. This review presents a comprehensive synthesis of the molecular pathways that regulate exosomal heterogeneity, with a focus on how these mechanisms govern the selective enrichment of biologically active cargo. We also discuss state-of-the-art technologies and omics platforms—such as ExoCarta and machine learning–based classifiers—used to decode exosomal content for diagnostic purposes. The clinical relevance of exosomes is examined through their roles in liquid biopsies for cancer, metabolic, and neurodegenerative diseases, emphasizing biomarker performance in terms of sensitivity and specificity. Furthermore, we explore the therapeutic potential of both native and engineered exosomes for targeted drug delivery, wound healing, and neuroregeneration, and provide insights into ongoing clinical trials. Despite growing interest, translational challenges persist, including standardization of isolation protocols, variability in cargo, targeting specificity, and regulatory constraints. Emerging strategies—ranging from synthetic exosome mimetics to AI-driven diagnostic algorithms—are rapidly reshaping the landscape of exosome-based precision medicine. This review consolidates current knowledge while proposing a forward-looking framework that integrates basic biology, engineering innovation, and clinical application, positioning exosomes as powerful agents in the future of diagnostics and therapeutics.

Introduction

Exosomes are small extracellular vehicles (EVs), typically 30–150 nm in diameter, released by virtually all cell types and are abundant across a spectrum of biofluids (e.g., blood, urine, cerebrospinal fluid) through the fusion of multivesicular bodies (MVBs) with the plasma membrane [1]. Originally dismissed as cellular debris, exosomes are now recognized as vital mediators of intercellular communication, transporting diverse bioactive molecules—proteins, lipids, nucleic acids, and metabolites—across local and systemic environments [1].

The cargo encapsulated within exosomes is not random. Instead, it reflects regulated packaging that can dramatically influence recipient cells. For instance, exosomes contain growth factors, cytokines, enzymes, membrane receptors, and transcription factors, enabling modulation of complex signaling networks. Lipid components such as ceramides, cholesterol, sphingomyelin, and lipid raft constituents constitute an integral part of the vesicle structure, and their specific properties and distribution play crucial roles in membrane curvature and in cargo sorting, forming and regulating these transport structures (see Table 1) [2]. Importantly, nucleic acids—including messenger RNAs (mRNAs), microRNAs (miRNAs), long non-coding RNAs (lncRNAs), circular RNAs (circRNAs), PIWI-interacting RNAs (piRNAs), and even DNA fragments—are selectively loaded, offering paracrine and potentially endocrine control of gene expression in target cells [3].

Recent studies underscore that exosomal cargo is shaped by both the physiological state and the origin of the donor cell. Cancer cell-derived exosomes, for example, often express oncogenic miRNAs (“oncomiRs”), mutated proteins, and immunomodulatory molecules, enabling them to promote tumor growth, angiogenesis, immune evasion, and metastasis [4]. Conversely, stem cell–derived exosomes are rich in regenerative signals like pro-angiogenic miRNAs and growth factors, contributing to tissue repair, anti-inflammation, and regenerative medicine applications [5]. Even exosomes from non-mammalian sources, including plants and bacteria, can carry functional miRNAs and lipids that modulate host immune responses and inflammatory pathways [6]. These findings underscore the emerging concept of cross-kingdom communication and broaden the translational relevance of extracellular vesicles beyond mammalian systems. To fully appreciate how such cross-kingdom vesicular signaling is achieved, it is essential to understand the molecular machinery governing exosome formation and cargo selection, which determines the specificity and functional capacity of these vesicles across biological systems.

The selective loading of bioactive molecules into exosomes is orchestrated by multiple, intersecting mechanisms. The canonical ESCRT (endosomal sorting complex required for transport) pathway involves ESCRT-0, –I, –II, –III, and associated proteins like ALIX, TSG101, and VPS4, which recognize ubiquitinated proteins and drive intraluminal vesicle (ILV) formation [7]. Alternative ESCRT-independent pathways include ceramide-mediated inward budding via neutral sphingomyelinase 2 (nSMase2), lipid rafts, and tetraspanin-enriched microdomains [6].

RNA cargo loading is equally multifaceted. Specific sequence motifs (e.g., the GGAG “EXOmotif”) in miRNAs are recognized by RNA-binding proteins (RBPs) such as hnRNPA2B1 or SYNCRIP; other RBPs include YBX1, Ago2, and MVP. MiRNA–RNA-induced silencing complex (miRISC)-dependent and lipid-raft–related mechanisms also contribute to cargo entry [6]. Ultimately, these sorting systems ensure exosomal cargoes are selectively packaged based on functional significance rather than passive encapsulation [8–10].

The selective and stable packaging of bioactive molecules renders exosomes valuable clinical biomarkers. Present in blood, urine, cerebrospinal fluid, saliva, breast milk, and other biofluids, they provide a non-invasive snapshot of donor cell status [1]. Exosomal miRNAs, proteins, and DNA have been identified as promising biomarkers for cancers (e.g., prostate, breast, glioblastoma), neurodegenerative diseases, cardiovascular disorders, and metabolic conditions. Their remarkable stability—due to vesicular protection—facilitates reliable measurement in clinical assays and the development of liquid biopsies to monitor disease progression and treatment response (Fig. 2).

Fig. 2.

Exosome biogenesis pathways - The schematic illustrates both ESCRT-dependent and ESCRT-independent mechanisms of intraluminal vesicle (ILV) formation within multivesicular bodies (MVBs). The ESCRT pathway (left) proceeds through sequential action of ESCRT-0, -I, -II, and -III complexes, with accessory proteins such as ALIX and TSG101. ESCRT-independent mechanisms (right) include ceramide-mediated budding via nSMase2, tetraspanin-enriched microdomains (CD9, CD63, CD81), and Rab GTPase-mediated regulation. These pathways may function independently or cooperatively, resulting in heterogeneous exosome populations with distinct cargo profiles

Therapeutically, exosomes hold vast promise owing to their innate delivery properties: biocompatibility, low immunogenicity, natural targeting, and ability to traverse physiological barriers, including the blood-brain barrier [11]. In cancer therapy, they have been engineered or loaded with chemotherapeutics (e.g., doxorubicin, paclitaxel), RNA therapeutics (miRNAs, siRNAs), and protein cargos to enhance targeted tumor delivery and reduce systemic toxicity. Similarly, mesenchymal stem cell–derived exosomes have been used to promote cardiac repair after myocardial infarction and enhance bone regeneration in preclinical models. Engineered exosomes—either pre-loaded via donor cell modification or post-loaded after isolation—offer programmable platforms with enhanced stability, extended circulation, and targeted delivery [7]. Others leverage surface engineering to display targeting ligands like peptides or antibodies, further fine-tuning therapeutic distribution.

Despite progress, several limitations temper clinical translation. Cargo heterogeneity, inefficient loading, variability in yield and purity, and challenges in scalable manufacturing and standardization remain unresolved. Immunogenicity concerns, potential horizontal transfer of unwanted nucleic acids (e.g., oncogenes), and regulatory hurdles (e.g., characterization, release criteria) must also be addressed. Importantly, rigorous clinical trial data are sparse; while early-phase trials exist, phase-III randomized studies demonstrating safety and efficacy are still pending.

This review aims to provide a thorough narrative of exosomal cargo—from the molecular mechanisms governing their selection, through advanced profiling techniques, to diagnostic and therapeutic applications, incorporating cutting-edge case studies (e.g., plant- and bacterial-derived exosomes) and critically assessing translational barriers.

The novelty of this review lies in its integrative synthesis of mechanistic and translational insights. While previous works have examined exosome biogenesis or clinical applications in isolation, this approach bridges fundamental molecular pathways of cargo regulation (ESCRT vs. non-ESCRT, RNA-binding proteins, lipid domains, PTMs) with their functional heterogeneity and clinical implications in diagnostics and therapeutics. By connecting these layers, this review highlights how molecular determinants of exosomal heterogeneity directly shape their diagnostic accuracy, therapeutic safety, and translational potential in precision medicine.

Exosomal biogenesis pathway

Exosomal biogenesis is a highly orchestrated, intracellular, multi-step process that converts intracellular components (endosomal cargo) into secreted nanovesicles, integrating canonical pathways, lipid–protein dynamics, and membrane trafficking regulators (Fig. 1). It entails the generation of intraluminal vesicles (ILVs) within multivesicular bodies (MVBs) and their eventual release into the extracellular space—a process regulated by both canonical and non-canonical pathways.

Fig. 1.

Exosomal Biogenesis Pathway - The diagram shows ESCRT-dependent pathways involving ESCRT-0, -I, -II, and -III complexes with ALIX and TSG101, and ESCRT-independent mechanisms such as ceramide-driven budding and tetraspanin-enriched microdomains. Together, these pathways enable selective cargo sorting and generate heterogeneous exosome populations with distinct biological functions

Intracellular origin: ILV formation & MVB fate

Exosomes form when the limiting membrane of late endosomes invaginates to generate ILVs, yielding exosomes (EVs) that range from 30 to 150 nm in diameter [12]. MVBs may either fuse with lysosomes or migrate to the plasma membrane to release ILVs as exosomes [12]. This mechanism differs from that of microvesicles (ectosomes) and apoptotic bodies, which originate by direct outward budding of the plasma membrane.

ESCRT vs. ESCRT-independent pathways

ESCRT-dependent and ESCRT-independent mechanisms operate through distinct molecular machineries and cellular contexts [3, 13]. ESCRT-dependent pathways involve sequential recruitment of ESCRT-0, -I, -II, and -III complexes, together with accessory proteins such as ALIX and TSG101, which coordinate cargo recognition, budding, and membrane scission [14]. This route predominates in physiological contexts such as immune surveillance and epithelial renewal, where ubiquitination tightly regulates vesicle sorting [15]. Conversely, ESCRT-independent mechanisms rely on lipid remodeling and microdomain organization. Ceramide generated by nSMase2 induces negative membrane curvature, tetraspanins (CD9, CD63, CD81) form sorting hubs, and Rab GTPases regulate vesicle trafficking [16, 17]. These lipid-driven pathways dominate in stress and oncogenic contexts—hypoxia, acidosis, and metabolic rewiring—where membrane fluidity and altered lipid composition favor vesicle release [18]. Together, these mechanisms ensure functional redundancy while tailoring exosomal cargo to cellular states.

Exosome biogenesis involves ESCRT-mediated pathways that require ESCRT-0, -I, -II, and -III complexes, together with essential accessory proteins including: ALIX, STAM1/2, HRS, TSG101, CHMP family protein, and ATPase VPS4, Rab GTPase/SNAREs. In contrast, non-ESCRT pathways depend predominantly on lipid composition (ceramide/sphingolipid metabolism) generated via nSMase2, tetraspanin-enriched microdomains such as CD9, CD63, and CD81, and Rab GTPases. Exosomal nucleic acid cargo is heterogeneous, containing diverse nucleic acids—including genetic material commonly found in the cytoplasm or nucleus: messenger RNAs (mRNAs), microRNAs (miRNAs), piwi-interacting RNAs (piRNAs), long non-coding RNAs (lncRNAs), small nucleolar RNAs (snoRNAs), small nuclear RNAs (snRNAs), circular RNAs (circRNAs), and genomic/mitochondrial DNA fragments—which are selectively recruited through RNA-binding proteins (e.g., hnRNPA2B1, YBX-1, SYNCRIP, Ago2) that recognize specific sorting motifs, along with lipid microdomain association [19].

ESCRT-Dependent pathways

The ESCRT (Endosomal Sorting Complexes Required for Transport) machinery constitutes the canonical pathway (Fig. 1) for intraluminal vesicle (ILV) formation within multivesicular bodies (MVBs), and it remains central to exosome biogenesis [20]. This pathway is orchestrated through a highly coordinated series of modular protein complexes—namely ESCRT-0, ESCRT-I, ESCRT-II, and ESCRT-III—as well as the ATPase VPS4, which is essential for component recycling. Accessory proteins such as ALIX (ALG-2-interacting protein X) and TSG101 (Tumor Susceptibility Gene 101) provide additional scaffolding and regulatory roles within this system.

ESCRT-0, composed primarily of HRS and STAM proteins, initiates the process by recognizing and clustering ubiquitinated transmembrane cargo on phosphatidylinositol 3-phosphate (PI3P)-enriched endosomal membranes [21]. This recognition event facilitates the sequential recruitment of ESCRT-I and ESCRT-II, which serve as structural adaptors that scaffold the membrane and drive initial budding processes [22]. The final step in membrane invagination and vesicle scission is executed by ESCRT-III, a dynamic assembly of charged multivesicular body proteins (CHMPs) that mediates membrane constriction. Following vesicle budding, the AAA ATPase VPS4 disassembles ESCRT-III polymers and recycles them for subsequent rounds of vesicle formation [23].

A noncanonical but functionally integrated route involves the syndecan–syntenin–ALIX axis, which directly recruits ESCRT-III in the absence of ESCRT-0, I, and II, thus facilitating ILV formation and cargo sorting [24]. The ALIX-mediated pathway contributes significantly to ILV formation under specific cellular conditions. Accessory ESCRT-III components—including CHMP1A, CHMP1B, CHMP5, and IST1—also promote ILV biogenesis within Rab11a-positive recycling endosomes, particularly during stress conditions such as nutrient deprivation [25, 26]. These findings suggest a stress-adaptive arm of the ESCRT pathway with potential significance for selective cargo packaging during cellular perturbations.

Despite the centrality of the ESCRT machinery, studies utilizing genetic knockdown or pharmacological inhibition of ESCRT components consistently show that exosome release is diminished but not entirely abolished [27]. This partial dependency underscores the existence of redundant and cooperative mechanisms, particularly involving lipid- and tetraspanin-mediated pathways, that ensure the robustness of exosome biogenesis even under compromised conditions. Thus, the ESCRT-dependent mechanism operates within a broader, modular system of vesicle formation that integrates multiple parallel routes to maintain vesicle heterogeneity and functional plasticity [28].

ESCRT-Independent pathways

While the ESCRT machinery remains a principal mechanism for ILV formation and exosome biogenesis, mounting evidence over the past decade has underscored the presence of ESCRT-independent pathways that act either complementarily or independently to ensure exosome heterogeneity and resilience in cargo delivery (Fig. 2) [29]. These alternative pathways largely rely on lipid-driven mechanisms, tetraspanin-enriched microdomains, and Rab GTPase-mediated regulation. The interplay between these pathways and canonical ESCRT-dependent routes is essential for functional plasticity, stress responsiveness, and selective cargo loading of exosomes [28].

Ceramide-driven budding

ESCRT-independent sorting is also driven by lipid microdomains. Ceramide, produced by sphingomyelinase, promotes negative curvature and intraluminal budding [30]. Importantly, lipid composition not only governs budding but also shapes tropism—cholesterol-rich exosomes preferentially target endothelial cells, while phosphatidylserine-enriched vesicles are more readily taken up by macrophages [31].

Among the most well-characterized ESCRT-independent mechanisms is the ceramide-mediated budding pathway. Ceramide, a bioactive sphingolipid, is generated through the hydrolysis of sphingomyelin by neutral sphingomyelinase 2 (nSMase2). This enzymatic activity induces spontaneous negative membrane curvature due to ceramide’s conical geometry, thereby promoting inward budding of endosomal membranes to form ILVs independent of protein scaffolds [32]. Ceramide-rich ILVs are more resistant to lysosomal degradation and are preferentially secreted as exosomes [33]. Inhibition of nSMase2 via pharmacological agents like GW4869 has been shown to significantly reduce exosome production in diverse cell types, confirming ceramide’s indispensable role [34].

Recent research has expanded the scope of ceramide involvement by highlighting the role of CERT (ceramide transfer protein), which facilitates ceramide transport from the endoplasmic reticulum (ER) to late endosomes through interactions with ESCRT-associated proteins like TSG101 [35]. This crosstalk between lipid-based and ESCRT-related pathways suggests a convergent regulatory landscape for ILV formation, offering cells a versatile response system to intracellular and extracellular cues.

Tetraspanin-Enriched microdomains (TEMs)

Tetraspanin-enriched microdomains act as molecular scaffolds, clustering proteins, lipids, and nucleic acids [36]. These processes are especially pronounced in cancer and hypoxia, where altered lipid metabolism enhances ceramide production and tetraspanin clustering. Tetraspanins—including CD9, CD63, CD81, and CD82—form cholesterol- and glycosphingolipid-enriched membrane platforms known as tetraspanin-enriched microdomains (TEMs). These microdomains act as organizational hubs for cargo selection and membrane remodeling, facilitating ILV budding in an ESCRT-independent manner [17]. TEMs recruit specific proteins such as PMEL, β-catenin, and Wnt components, and serve as scaffolds for selective loading of signaling molecules, RNAs, and lipids [37]. Tetraspanins are also critical for membrane curvature and fission through their lateral interactions with lipid rafts and cytoskeletal elements.

Recent insights suggest that tetraspanins may act as compensatory mechanisms when ESCRT components are deficient or genetically ablated. For example, CD81 and CD9 have been shown to facilitate exosome release in cells lacking functional ESCRT proteins, supporting their autonomous capacity to drive ILV formation [38, 39]. Additionally, post-translational modifications such as palmitoylation have been identified as regulatory switches that control tetraspanin localization and clustering during ILV formation [40].

Rab31- and Lipid-mediated mechanisms

An emerging noncanonical regulator of ILV formation is Rab31 (also known as Rab22B), a small GTPase that promotes exosome biogenesis through interaction with flotillin-1 and LAMP2A, both of which localize to cholesterol-rich microdomains. Rab31 not only facilitates ILV budding but also diverts cargo away from lysosomal degradation, thereby enhancing exosome release [41]. This ESCRT-independent route appears to be particularly relevant in cancer cells, where elevated Rab31 expression correlates with aggressive metastatic phenotypes and altered exosomal signaling [41]. Moreover, lipid mediators such as sphingosine-1-phosphate (S1P) [42] and ceramide [43] have been shown to modulate exosome biogenesis through direct effects on tetraspanin sorting and endosomal membrane dynamics. S1P, in particular, regulates calcium-dependent exocytosis and enhances MVB–plasma membrane fusion, suggesting its dual role in both ILV formation and exosome secretion [44]. These lipid-driven processes provide an additional layer of regulation and confer responsiveness to metabolic and inflammatory states.

Integration of ESCRT-independent pathways

The ESCRT-independent mechanisms—though individually distinct—often intersect with one another and with ESCRT-dependent processes to produce functionally distinct exosome subtypes. For instance, ceramide and tetraspanin microdomains have been implicated in the selective loading of microRNAs and signaling lipids, while Rab31 pathways specialize in redirecting endosomal traffic away from lysosomal degradation [29]. These parallel pathways ensure the generation of heterogeneous exosome populations tailored to the physiological and pathological needs of the cell. Studies using high-resolution single-vesicle imaging and proteomics have revealed that ESCRT-independent exosomes often differ in cargo composition, surface markers, and bioactivity compared to their ESCRT-dependent counterparts [45]. This emerging complexity highlights the need for more refined classification systems and emphasizes the importance of understanding ESCRT-independent pathways in developing targeted exosome-based therapies.

Cargo recognition and sorting signals

The process of cargo recognition and selective loading into intraluminal vesicles (ILVs) within multivesicular bodies (MVBs) is a tightly regulated event that underpins the biological specificity of exosome-mediated intercellular communication. Unlike passive diffusion, cargo sorting involves a suite of molecular signals, including post-translational modifications, lipid interactions, and RNA–protein complexes that coordinate the recruitment of biomolecules into budding vesicles. These mechanisms operate in parallel and often intersect between ESCRT-dependent and independent pathways, contributing to the complexity and heterogeneity of the exosomal cargoome [39, 46].

Ubiquitination and ESCRT-mediated targeting

Ubiquitination serves as a canonical signal for the inclusion of proteins into ILVs via the ESCRT machinery. Specifically, ESCRT-0 components such as HRS and STAM recognize monoubiquitinated cargo on PI3P-enriched endosomal membranes (Table 1), facilitating their downstream handoff to ESCRT-I and ESCRT-II for vesicle budding [47]. This hierarchical recognition ensures specificity in the incorporation of signaling receptors, ion channels, and transporters into exosomes. Moreover, deubiquitinating enzymes such as USP8 and AMSH modulate this process, suggesting dynamic control over cargo sorting [48].

Table 1.

Representative RNA-binding proteins (RBPs) and regulatory mechanisms involved in the selective loading of RNA cargo into exosomes

Sorting Signal	Mechanism	Key Molecules	Pathway	References
Ubiquitination	Tagged cargo recognized by ESCRT-0 (HRS/STAM)	HRS, TSG101, ALIX	ESCRT-dependent	[47, 48]
EXOmotifs in miRNA	Binding to RBPs for selective RNA loading	hnRNPA2B1, SYNCRIP, YBX-1, Ago2	ESCRT-dependent & ESCRT-independent	[46]
PTMs (e.g., SUMOylation)	Modulate cargo affinity for sorting machinery	α-synuclein, tau proteins	ESCRT-dependent & ESCRT-independent	[49]
Lipid Microdomains	Provide curvature and a platform for ceramide-mediated budding	Ceramide, sphingomyelin, and cholesterol	ESCRT-independent	[30]
Lipid Transfer Proteins	Coordinate the delivery of lipids that initiate budding	CERT, ORP1L, LAMP2A	ESCRT-independent	[50]

RNA motifs and RNA-Binding proteins (RBPs)

Selective RNA loading into exosomes is orchestrated by RNA-binding proteins (RBPs) that recognize specific sequences or structural motifs [7, 51]. hnRNPA2B1 binds EXOmotifs in miRNAs in a SUMOylation-dependent manner, ensuring preferential loading of oncogenic and stress-response miRNAs. SYNCRIP recognizes hEXO motifs, particularly regulating hepatocyte exosomal RNA content involved in lipid metabolism. YBX-1 mediates selective packaging of small RNAs, including oncogenic miR-223, and is critical under conditions of cellular stress [10, 46]. Collectively, these RBPs act as molecular gatekeepers that dynamically reprogram exosomal RNA signatures in response to the cellular environment, underpinning their disease specificity [52]. Selective loading of RNA, particularly microRNAs (miRNAs) and messenger RNAs (mRNAs), is mediated by conserved sequence motifs and their interaction with RBPs. The GGAG EXOmotif, found in several exosome-enriched miRNAs, is recognized by hnRNPA2B1, which undergoes SUMOylation to facilitate its function in sorting [46]. Other key RBPs include SYNCRIP, YBX-1, Ago2, and FMR1, each contributing to the selective inclusion of RNA subsets depending on the cellular context [53, 54]. Emerging evidence has also implicated noncoding RNAs and RNA modifications such as N6-methyladenosine (m6A) in influencing RNA–RBP interactions and sorting, expanding the regulatory repertoire of exosomal RNA cargo [55].

Post-Translational modifications (PTMs)

Post-translational modifications (PTMs) serve as molecular switches that regulate cargo recruitment and vesicle function [49]. SUMOylation of hnRNPA2B1 enhances recognition of miRNA EXOmotifs, phosphorylation of tau directs its sorting into neuronal exosomes in Alzheimer’s disease, and glycosylation of mucins alters exosomal immunogenicity in cancer [46]. These modifications produce disease-specific vesicle signatures, explaining how distinct exosome populations emerge in cancer versus neurodegeneration, and offering potential for biomarker discovery. PTMs such as phosphorylation, SUMOylation, ISGylation, and glycosylation modulate the inclusion of proteins into ILVs by altering their affinity for sorting machinery [56] (Table 1). For instance, SUMOylated α-synuclein and phosphorylated tau proteins are preferentially sorted into exosomes in neurodegenerative models, implicating PTMs in disease-associated vesicle populations [49]. Additionally, ISGylation—an interferon-induced ubiquitin-like modification—has been shown to control cargo packaging in response to viral infections, highlighting the context-dependent nature of exosome biogenesis [57].

Lipid affinity and membrane microdomains

Lipid composition of the endosomal membrane plays a critical role in cargo sorting. Ceramide, sphingomyelin, and cholesterol-rich microdomains act as platforms for selective cargo inclusion, especially in ESCRT-independent pathways (Table 1). For example, the sorting of proteolipid protein (PLP) into ILVs has been shown to depend on ceramide-rich domains, linking lipid affinity with membrane curvature and cargo capture [30]. The involvement of lipid transfer proteins such as CERT and ORP1L, as well as cholesterol transporters, further modulates the formation of lipid microdomains conducive to cargo sequestration [50].

Convergence and subpopulation formation

Notably, these cargo-sorting mechanisms are not mutually exclusive. Cells frequently employ overlapping cues—ubiquitination for protein selection, EXOmotifs for RNA targeting, and lipid microdomain localization—to achieve precise and context-specific exosomal outputs (Table 1). As a result, individual MVBs may generate ILVs that differ in cargo composition despite originating from the same cell, underpinning the functional heterogeneity of exosome populations [45]. This multiplicity allows cells to tailor their exosomal output to environmental cues such as stress, inflammation, hypoxia, or oncogenic transformation, thus making exosomes dynamic conveyors of cellular state.

Illustration of parallel and intersecting mechanisms of exosomal cargo sorting. Both ESCRT-dependent and -independent routes leverage ubiquitination, RNA–RBP interactions, lipid microdomains, and PTMs to ensure selective cargo loading and ILV heterogeneity (Fig. 2).

Functional heterogeneity of Exosomal bioactive molecules

Exosomes exhibit remarkable functional heterogeneity, reflecting the physiological state, stress response, and microenvironment of their parent cells. This diversity allows exosomes to modulate a wide range of biological processes—from immune regulation and angiogenesis to cellular reprogramming and tumor progression. Exosomes carry a complex cargo of bioactive molecules—including proteins, lipids, nucleic acids, and metabolites—which contribute to their versatile roles in intercellular communication, immune modulation, disease propagation, and therapeutic regulation (Fig. 3). These biomolecules are not merely passively loaded but are selectively sorted to reflect the physiological or pathological state of the parent cell, endowing exosomes with functional specificity in local and systemic environments. Tumor-derived exosomes are enriched in oncogenic miRNAs, mutated and immunomodulatory proteins, and metastasis-promoting factors that enhance angiogenesis, invasion, immune evasion, and metabolic reprogramming [58]. In contrast, stem-cell–derived exosomes predominantly contain regenerative and pro-repair cargos, including proangiogenic miRNAs, growth factors, and anti-inflammatory mediators that support tissue healing and immune regulation [59].

Fig. 3.

Clinical relevance of exosomal cargo - The figure is a conceptual schematic illustrating exosomal cargo sorting as both a diagnostic and therapeutic target, emphasizing how regulated cargo selection underlies exosomal functional heterogeneity. Selective exosomal cargo sorting, where proteins, lipids, nucleic acids, and metabolites reflect the physiological state and microenvironment of parent cells. The regulated heterogeneity underpins exosome-mediated immune modulation, angiogenesis, cellular reprogramming, tumor progression, and therapeutic responses

Exosomal proteins: structural and functional mediators

Beyond RNA, exosomes encapsulate nuclear and mitochondrial DNA, which mirror the mutational landscape of parent cells [60, 61]. Exosomal proteins serve both structural and functional purposes. Structural proteins such as tetraspanins (CD63, CD81, CD9), heat shock proteins (HSP70, HSP90), and cytoskeletal elements (actin, tubulin) facilitate vesicle integrity and trafficking. Functionally, exosomal proteins can regulate immune responses, angiogenesis, tumor progression, and organotropism [62]. For instance, exosomal ALIX and TSG101—components of the ESCRT complex—regulate ILV formation and are often used as exosome markers, while also participating in protein sorting [29]. Cancer-derived exosomes have been found enriched in metalloproteinases (e.g., MMP-9), which facilitate extracellular matrix remodeling and metastatic invasion [63]. Furthermore, immune regulatory molecules such as PD-L1 expressed on exosomes can suppress T-cell activity and contribute to immune evasion in cancers [64].

Functional roles – immune modulation

Exosomes act as critical mediators of immune regulation by transferring antigens, cytokines, and immunomodulatory molecules. Dendritic cell–derived exosomes, for instance, stimulate T-cell responses by presenting MHC–peptide complexes, while tumor-derived exosomes enriched in PD-L1 suppress cytotoxic T lymphocyte activity [65, 66]. Failures in this balance are linked to disease: excessive release of immunosuppressive exosomes in cancer promotes immune escape, whereas impaired exosome-mediated antigen presentation in autoimmune disease can blunt tolerance [67–69]. Thus, exosomal interactions with the immune system can either sustain immune surveillance or contribute to immune evasion, depending on their cellular origin and cargo. Exosomes exert dual immunoregulatory effects: dendritic cell-derived exosomes can enhance antigen presentation and stimulate T-cell activation, whereas tumor- or MSC-derived exosomes may suppress immune responses by delivering PD-L1, FasL, or immunosuppressive cytokines to inhibit cytotoxic immunity [70, 71].

Nucleic acids: miRNAs, lncRNAs, and mRNAs in Exosomal communication

Exosomal DNA provides a stable, protected substrate for minimally invasive genomic profiling in liquid biopsy applications [72–74]. Moreover, double-stranded DNA in exosomes can activate innate immune pathways such as cGAS–STING, linking vesicle cargo to immunomodulation in cancer and autoimmunity. Recent findings indicate that the immunogenic potential of exosomal DNA is influenced by both its cellular origin and structural characteristics. Exosomes derived from cancer cells often contain fragmented double-stranded DNA (dsDNA) and micronuclei-derived DNA, which act as potent activators of the cGAS–STING signaling cascade, stimulating type I interferon responses and proinflammatory signaling [75]. In contrast, exosomes from healthy cells typically harbor shorter or protein-bound DNA fragments that limit cGAS access and attenuate innate immune activation. The size, methylation status, and chromatin association of these DNA fragments collectively determine their ability to activate cGAS and STING, explaining why tumor-derived exosomal DNA exhibits greater immunostimulatory potency than DNA from non-malignant sources. Exosomes are rich in nucleic acids, including messenger RNAs (mRNAs), microRNAs (miRNAs), long non-coding RNAs (lncRNAs), and circular RNAs (circRNAs), many of which are selectively enriched via RNA-binding proteins (RBPs) like hnRNPA2B1, SYNCRIP, and YBX-1 [46, 76]. These RNAs can function in recipient cells to modulate gene expression, promote epigenetic remodeling, and initiate signaling cascades. In cancer, exosomal miR-21 and miR-155 are well-characterized for their roles in promoting tumor growth, angiogenesis, and immune suppression [77]. Similarly, lncRNAs such as MALAT1 and HOTAIR in exosomes derived from lung or breast cancer cells have been implicated in pre-metastatic niche formation and epithelial-mesenchymal transition (EMT) [78, 79]. RNA cargo can also serve as disease biomarkers. For example, exosomal miR-1246 has shown diagnostic promise in colorectal cancer, while exosomal lncRNA GAS5 has been linked to breast cancer prognosis [80].

Exosomal lncRNA GAS5 has been implicated functionally as a tumor-suppressive regulator rather than a passive biomarker. Mechanistically, GAS5 acts as a competing endogenous RNA (ceRNA) that sponges oncogenic miRNAs (thereby derepressing tumor-suppressor targets), modulates apoptotic and cell-cycle pathways through interactions with nuclear receptors (notably the glucocorticoid receptor), and influences key survival pathways including mTOR and PI3K/AKT [81]. When exported within exosomes, GAS5 can be transferred to recipient cells where it suppresses proliferation, reduces invasive behavior, and increases chemosensitivity, likely by restoring expression of pro-apoptotic and cell-cycle regulatory genes. A recent comprehensive review synthesizes these mechanistic roles and highlights how multi-omics profiling combined with machine-learning approaches can validate GAS5-centered regulatory networks and strengthen its prognostic utility in breast cancer [82]. Together, these data support a model in which exosomal GAS5 functions both as a mechanistic mediator of anti-tumor signaling and as a robust candidate biomarker for prognosis and therapy response.

Lipids: structural scaffolds and signaling mediators

Exosomes are enriched in bioactive lipids, including sphingolipids, ceramides, cholesterol, and phosphatidylserine. These lipids not only maintain membrane curvature and vesicle stability but also participate in signaling pathways. Ceramide, generated via neutral sphingomyelinase (nSMase2), facilitates ILV budding and exosome release through membrane invagination [30]. Lipid mediators such as prostaglandins, leukotrienes, and sphingosine-1-phosphate (S1P) within exosomes contribute to inflammatory regulation and immune cell recruitment [83, 84]. Recent studies show that lipid compositions vary across cell types and disease states, influencing exosomal tropism and uptake efficiency [85]. Lipidomics has revealed that tumor-derived exosomes possess elevated cholesterol and saturated fatty acids, modulating membrane fusion and cargo delivery [86].

Metabolites and enzymes: metabolic regulation via Exosomal transfer

Although underexplored compared to nucleic acids and proteins, emerging evidence supports the presence of metabolic enzymes and small metabolites in exosomes. These components can reprogram cellular metabolism in recipient cells, facilitating tumor proliferation or immune suppression. For example, exosomes from hypoxic tumor cells carry lactate dehydrogenase (LDH), contributing to the Warburg effect and promoting angiogenesis [87]. Tumor-derived exosomes also transfer glycolytic enzymes such as pyruvate kinase M2 (PKM2) and enolase-1, enhancing aerobic glycolysis in stromal cells [88, 89]. Metabolomic profiling has detected amino acids, TCA cycle intermediates, and lipids in plasma exosomes, suggesting systemic metabolic influence, with potential applications in metabolic syndrome and cancer cachexia diagnostics [90].

Functional implications across biological systems

The heterogeneity of exosomal bioactive molecules allows them to influence diverse physiological and pathological processes across organ systems. By transferring context-specific molecular cargo—such as proteins, RNAs, lipids, and metabolites—exosomes act as dynamic mediators of cell-to-cell communication that fine-tune immune, neural, metabolic, and oncogenic signaling networks. Cancer: Tumor-derived exosomes play a pivotal role in shaping the tumor microenvironment and promoting malignancy. They transfer oncogenic proteins, microRNAs, and metabolites that support tumor cell proliferation, invasion, angiogenesis, and metastasis. Moreover, they facilitate immune escape by carrying checkpoint molecules such as PD-L1 and suppressing cytotoxic T-cell responses. Exosomes also contribute to chemoresistance by exporting drug efflux transporters or anti-apoptotic factors to recipient cells, thereby enhancing tumor survival under therapeutic pressure. Immunology: In the immune system, exosomes function as both activators and suppressors of immune responses. Dendritic-cell–derived exosomes can present antigens via MHC-peptide complexes, stimulating T-cell activation and adaptive immunity. Conversely, tumor or mesenchymal stem cell (MSC)-derived exosomes may exert immunosuppressive effects, dampening macrophage and T-cell responses through cytokines, FasL, or TGF-β signaling. This dual role underscores their context-dependent immunomodulatory capacity in inflammation, infection, and cancer immunotherapy. Neurobiology: In the central nervous system, exosomes act as crucial vehicles for neuron–glia communication and synaptic maintenance. They carry signaling molecules that regulate synaptic plasticity, axonal growth, and neuroprotection, but also mediate the spread of neurotoxic proteins such as tau and amyloid-β in Alzheimer’s disease, or α-synuclein in Parkinson’s disease [91]. Thus, exosomes represent both physiological messengers and pathological propagators within neural networks. Exosomes facilitate the spread of neurodegenerative pathology by transporting misfolded proteins—such as tau and α-synuclein—between neural cells, thereby promoting the propagation of Alzheimer’s and Parkinson’s disease–associated aggregates. Cardiovascular Disease: Exosomes are increasingly recognized as regulators of vascular remodeling and cardiac repair. Endothelial and cardiac cell-derived exosomes enriched in miR-21, miR-126, and miR-210 modulate fibrosis, angiogenesis, and vascular smooth muscle proliferation, contributing to the remodeling processes underlying atherosclerosis, myocardial infarction, and heart failure. Depending on their source, they can either exacerbate pathological remodeling or facilitate tissue regeneration following injury. Infectious Disease: Pathogens exploit the exosomal pathway to enhance infection and immune evasion. Viruses such as HIV and SARS-CoV-2 incorporate viral RNA, proteins, or host-derived immune modulators into exosomes, enabling intercellular dissemination without direct viral exposure. These vesicles can suppress antiviral immune responses, manipulate cytokine signaling, and promote systemic inflammation. Understanding this interplay between pathogens and exosome biology provides valuable insight into infection dynamics and antiviral therapeutic design.

While functional diversity defines the biological outcomes of exosome signaling, this heterogeneity originates from tightly regulated molecular mechanisms that determine which biomolecules are packaged and secreted. The next section examines these molecular processes governing exosome biogenesis and selective cargo loading. Exosomal functional heterogeneity highlights their role as adaptable mediators of intercellular communication. Their ability to tailor biological responses across diverse systems underpins their emerging applications in diagnostics and therapeutics.

Molecular regulation of Exosomal bioactive molecules

The molecular mechanisms driving exosome diversity are multifactorial, involving both ESCRT-dependent and ESCRT-independent pathways, lipid microdomain dynamics, and a network of RNA-binding proteins and post-translational modifications (PTMs). These processes collectively dictate the biogenesis, cargo selection, and release of exosomes, shaping their compositional and functional heterogeneity. The molecular regulation of exosome biogenesis and cargo selection (Fig. 4) is a finely orchestrated process influenced by cell type, physiological state, and microenvironmental cues. This regulation underlies the heterogeneity of exosomes, reflected in their cargo composition, functional properties, and tropism. Understanding these regulatory mechanisms is crucial for harnessing exosomes in diagnostics and therapeutics.

Fig. 4.

Functional Heterogeneity of Exosomal Bioactive Molecules - Exosomes carry proteins, lipids, nucleic acids, DNA, and metabolites. “Nucleic acids” in this context refers specifically to RNA classes (mRNAs, miRNAs, lncRNAs, circRNAs), while “DNA” is highlighted separately to emphasize its distinct roles in diagnostics (liquid biopsies) and immune modulation

Biogenesis pathways drive Exosomal heterogeneity

Both ESCRT-dependent and ESCRT-independent pathways contribute to exosome biogenesis, but their relative dominance varies with cell type and physiological context. In most somatic and stem cells, the ESCRT machinery (comprising ESCRT-0, -I, -II, and -III complexes, along with accessory proteins like ALIX and VPS4) serves as the primary route for intraluminal vesicle (ILV) formation and cargo sequestration, reflecting its critical role in maintaining endosomal homeostasis and regulated vesicle turnover [47]. In contrast, many cancer cells exhibit enhanced reliance on ESCRT-independent mechanisms, particularly those mediated by neutral sphingomyelinase 2 (nSMase2) and ceramide-driven microdomains, which facilitate rapid vesicle budding under oncogenic or hypoxic stress. Tetraspanin-enriched microdomains (CD9, CD63, CD81) also contribute substantially to cargo selection and membrane curvature in these ESCRT-independent pathways, often compensating for downregulated or mutated ESCRT components [17, 32]. Evidence suggests that such pathway flexibility allows malignant cells to sustain high exosome output and adapt their secretome to environmental stress, whereas stem cell–derived exosomes maintain ESCRT dependence to ensure controlled cargo sorting and regenerative signaling fidelity [29, 30, 92, 93].

RNA-Binding proteins and motifs regulate selective RNA loading

Selective RNA loading into exosomes is orchestrated by RNA-binding proteins (RBPs) that recognize specific sequences or structural motifs. hnRNPA2B1 binds EXOmotifs in miRNAs in a SUMOylation-dependent manner, ensuring preferential loading of oncogenic and stress-response miRNAs [46]. SYNCRIP recognizes hEXO motifs, particularly regulating hepatocyte exosomal RNA content involved in lipid metabolism [76, 94]. YBX-1 mediates selective packaging of small RNAs, including oncogenic miR-223, and is critical under conditions of cellular stress (92]. Collectively, these RBPs act as molecular gatekeepers that dynamically reprogram exosomal RNA signatures in response to the cellular environment, underpinning their disease specificity.

The selective enrichment of RNAs into exosomes is a tightly regulated process controlled by specific sequence motifs and RNA-binding proteins (RBPs) that recognize and direct RNA cargo toward vesicular packaging. Among the best-characterized examples, hnRNPA2B1 has been shown to recognize specific GGAG motifs within miRNAs, facilitating their recruitment into exosomes through SUMOylation-dependent mechanisms [46]. This modification acts as a molecular switch, enhancing hnRNPA2B1’s affinity for target RNAs and ensuring selective cargo sorting rather than random encapsulation. Similarly, SYNCRIP and YBX-1 function as key adaptors that mediate the selective loading of miRNAs and mRNAs into exosomes [53]. YBX-1, for instance, binds to distinct sequence motifs within small RNAs and stabilizes their incorporation into vesicles, linking post-transcriptional regulation to extracellular communication. SYNCRIP, on the other hand, recognizes a conserved “hEXO motif” in miRNAs, coordinating vesicular export and maintaining cell-specific RNA signatures in the exosomal population. Beyond protein recognition, RNA chemical modifications such as N6-methyladenosine (m6A) play a crucial role in determining which transcripts are preferentially packaged. The presence of m6A marks modulates the interaction between RBPs and RNA substrates, effectively serving as molecular signals that dictate RNA export into vesicles [95]. Through these multi-layered regulatory processes—sequence recognition, RBP mediation, and epitranscriptomic modification—cells achieve a high level of selectivity and functional precision in exosomal RNA cargo composition.

The selective incorporation of RNA species into exosomes is orchestrated by a network of RNA-binding proteins (RBPs) that recognize distinct sequence motifs and structural elements on target RNAs. Rather than functioning in direct competition, these RBPs appear to operate in a context-dependent and partially hierarchical manner, adapting to cellular conditions and RNA availability. For example, hnRNPA2B1 recognizes GGAG motifs within specific miRNAs and facilitates their sorting into exosomes through SUMOylation-dependent modification, which enhances its RNA-binding affinity and exosomal association. In contrast, SYNCRIP (also known as hnRNP-Q) interacts with GGCU/A motifs through its N-terminal RNA recognition domains and primarily governs hepatocyte and immune-cell exosomal miRNA loading during inflammatory or stress responses. Experimental evidence suggests that when multiple RBPs target overlapping RNA subsets, their binding is temporally or spatially segregated—regulated by post-translational modifications (e.g., SUMOylation, phosphorylation) and subcellular localization within multivesicular bodies. Thus, rather than competing for identical RNA cargo, hnRNPA2B1, SYNCRIP, and related RBPs coordinate RNA selection through motif-specific, condition-sensitive, and possibly sequential mechanisms, maintaining the fidelity and adaptability of exosomal RNA composition [76, 94]. These RBPs act as molecular “gatekeepers,” determining which transcripts are sorted into exosomes versus retained intracellularly, thereby contributing to the functional specialization of exosomes.

Protein sorting and Post-Translational modifications

The sorting of proteins into exosomes is a highly orchestrated process governed by post-translational modifications (PTMs) such as ubiquitination, SUMOylation, glycosylation, and phosphorylation. These modifications act as molecular codes that influence protein–protein interactions and determine whether specific cargoes are directed toward inclusion within intraluminal vesicles (ILVs) through the ESCRT machinery or tetraspanin-enriched microdomains.

For instance, ubiquitination serves as a key signal for protein selection by the ESCRT-0 complex, particularly through the recognition of ubiquitinated substrates by HRS, which guides them toward endosomal sorting and eventual exosomal release [96]. This mechanism ensures the selective inclusion of signaling or membrane proteins and prevents indiscriminate trafficking. Similarly, glycosylation plays a pivotal role in defining the exosomal proteome, especially in cancer cells. For example, SUMOylated α-synuclein is selectively sorted into exosomes, and glycosylated forms of MUC1 in exosomes are associated with immune [97]. The glycosylation of integrins and mucins (such as MUC1) enhances their recruitment into tumor-derived exosomes, thereby promoting processes like metastasis and immune evasion, and chemoresistance in breast cancer [97] through improved adhesion, receptor binding, and shielding from immune recognition [97]. Another regulatory layer involves phosphorylation of chaperone proteins, notably heat shock proteins (HSPs), which alters their conformation and modulates both their chaperone activity and exosomal destination [98]. This modification influences how HSPs participate in protein folding, stress signaling, and immune modulation once exported in vesicles. Post-translational modifications function as molecular switches, fine-tuning cargo selection and shaping the functional identity of exosomes according to the cell’s physiological or pathological state.

These post-translational modifications (PTMs) increase cargo heterogeneity and reflect cellular signaling states, contributing to exosome heterogeneity. PTMs act as molecular switches in exosome biology. SUMOylation of hnRNPA2B1 enhances miRNA sorting, phosphorylation of tau influences its inclusion in neuronal exosomes, and glycosylation of mucins alters cancer exosomal immunogenicity [46]. These PTM-dependent signatures enable disease-specific vesicle populations, offering mechanistic insight into heterogeneity and biomarker specificity.

Lipid-Mediated regulation and membrane microdomains

Lipid composition is not only a structural determinant but also a regulatory factor in exosome formation. Ceramide, sphingosine-1-phosphate (S1P), and cholesterol enrich membrane curvature, facilitate budding, and sort specific proteins. Lipid-modifying enzymes such as nSMase2 are tightly regulated and sensitive to cellular stress, hypoxia, and inflammation—modulating exosomal output and content [99]. Thus, lipid-based regulation acts as a dynamic sensor of microenvironmental stimuli and influences exosomal phenotype. Lipid microdomains provide an ESCRT-independent mechanism of cargo selection. Ceramide, produced by nSMase2, promotes negative curvature of the endosomal membrane, driving budding and vesicle release [100]. Tetraspanin-enriched microdomains organize protein and RNA cargo, clustering functionally related molecules into specific vesicle subtypes [17, 101]. These lipid-driven processes are particularly upregulated in malignant cells, linking metabolic rewiring to exosomal heterogeneity.

Tumor microenvironment and pathological regulation

The tumor microenvironment (TME) exerts a profound influence on exosomal cargo composition. Hypoxia induces HIF-1α, enriching vesicles with angiogenic factors such as VEGF and miR-210 [102, 103]. Acidosis enhances the secretion of proteases and ECM-modifying enzymes [104]. Stromal interactions promote metabolic reprogramming, while immune suppression within the TME enriches vesicles with PD-L1 and immunomodulatory cytokines [105, 106]. The tumor microenvironment (TME)—characterized by hypoxia, acidosis, nutrient deprivation, and immune infiltration—exerts a profound influence on exosomal heterogeneity and function. These environmental pressures shape the molecular composition of exosomes, altering their roles in tumor progression, angiogenesis, and immune modulation. Under hypoxic conditions, stabilization of HIF-1α promotes the release of exosomes enriched with angiogenic factors such as VEGF and miR-210, which stimulate endothelial cell proliferation and new vessel formation [107]. These vesicles act as paracrine mediators, allowing tumors to adapt to oxygen-deprived niches and sustain their metabolic needs. Similarly, oncogenic mutations, including those in KRAS and TP53, profoundly remodel exosomal content [108]. Mutant KRAS drives the secretion of vesicles carrying pro-tumorigenic RNAs, DNA fragments, and signaling proteins that reinforce malignant phenotypes and enhance the invasive potential of neighboring cells. Mutant TP53, in turn, promotes the packaging of molecules that facilitate immune evasion and resistance to apoptosis. Furthermore, immune signaling within the TME regulates exosomal expression of PD-L1, ICAMs, and MHC molecules, thereby modulating the crosstalk between cancer and immune cells [64]. Exosomal PD-L1, for example, binds to PD-1 receptors on T cells, suppressing their cytotoxic function and enabling tumor immune escape. In parallel, alterations in ICAM and MHC expression contribute to impaired antigen presentation and metastasis.

In addition to hypoxia and oncogenic signaling, tumor acidosis plays a critical role in modulating exosomal secretion and function. Acidic tumor microenvironments enhance the release of exosomal proteases, particularly MMP-2 and MMP-9, which are key mediators of extracellular matrix (ECM) remodeling [109]. These proteases degrade structural ECM components such as collagen IV and fibronectin, loosening the intercellular scaffold and facilitating both tumor invasion and exosome internalization. The resulting ECM breakdown products can act as signaling cues, activating integrin- and CD44-dependent endocytic pathways that further promote vesicle uptake by neighboring cells [110]. This establishes a feedback loop in which ECM degradation accelerates exosome dissemination, while exosomal proteases reinforce matrix remodeling and metastatic spread [111]. Recent evidence indicates that this reciprocal interaction is especially pronounced under acidic and hypoxic conditions, contributing to tumor aggressiveness and stromal reprogramming [112]. Collectively, these TME-driven alterations give rise to highly heterogeneous exosome populations that dynamically orchestrate tumor angiogenesis, metastasis, and immune evasion, reinforcing the adaptability and resilience of cancer ecosystems. This contextual regulation underscores why exosome composition varies across cancer types, disease stages, and treatment responses, complicating the development of universal exosome-based biomarkers.

Building upon these regulatory insights, it becomes essential to understand the molecular composition of exosomes and how recent high-throughput technologies are profiling their complex cargo landscape. Section 5 provides an overview of these molecular profiling strategies and bioinformatic resources.

Molecular regulation orchestrates the generation of exosomal diversity, enabling cells to tailor the information carried within vesicles. Understanding these underlying mechanisms provides the molecular framework for interpreting the biological and diagnostic significance of exosomal cargo.

Molecular cargo profiling

Recent advances in high-throughput omics and bioinformatics have transformed our understanding of exosomal cargo composition. Proteomic, lipidomic, and transcriptomic profiling now reveals the complex molecular landscape of exosomes across cell types and disease states, enabling precise mapping of functional and diagnostic signatures. The molecular cargo of exosomes comprises a diverse array of bioactive molecules that reflect the state, type, and physiological context of the parent cell. Profiling these cargos (Fig. 5) has revealed their functional roles in intercellular communication, immune modulation, tumor progression, and tissue regeneration. Comprehensive molecular characterization of exosomal content—particularly proteins, lipids, nucleic acids, and metabolites—has become central to understanding their biology and clinical applications.

Fig. 5.

Molecular regulation of exosome biogenesis and cargo selection - Exosome heterogeneity arises from coordinated molecular pathways. The ESCRT machinery (HGS, TSG101, VPS4, STAM) governs membrane remodeling and cargo sequestration. RNA-binding proteins (RBPs) recognize sequence motifs to direct selective RNA loading. Lipid regulators, including ceramide and tetraspanin-rich domains, influence budding and packaging. Post-translational modifications (PTMs) act as molecular switches to modulate protein and RNA sorting. Together, these mechanisms shape the heterogeneity and functional specialization of exosomal populations

Exosomal cargo composition

Proteins

Exosomes carry a rich proteome, including both conserved and cell-type-specific proteins. Common markers such as tetraspanins (CD9, CD63, CD81), ALIX, and TSG101 are associated with exosomal membranes and used for their identification [39]. The proteomic profile also contains enzymes (e.g., GAPDH, PKM2), heat shock proteins (Hsp70, Hsp90), adhesion molecules, and cytoskeletal components, reflecting the parent cell’s state and origin [45]. Pathology-specific exosomal proteins, such as glypican-1 in pancreatic cancer [113] and PD-L1 in tumor immune evasion, have been identified as potential diagnostic or prognostic markers.

Nucleic acids

Exosomes encapsulate diverse nucleic acid species that play critical roles in mediating genetic exchange and post-transcriptional regulation in recipient cells. These nucleic acids contribute to intercellular communication, influencing gene expression, immune modulation, and disease progression. mRNAs: Functional messenger RNAs (mRNAs) packaged within exosomes can be transferred to target cells, where they are translated into functional proteins, thereby transmitting phenotypic traits or adaptive responses across cells [114]. This mechanism represents a form of horizontal information transfer that extends beyond conventional cell signaling. miRNAs: Exosomal microRNAs (miRNAs) act as potent post-transcriptional regulators, repressing or silencing target mRNAs in recipient cells. Notably, miR-21, miR-1246, and miR-155 are consistently enriched in tumor-derived exosomes, where they modulate pathways controlling cell proliferation, invasion, apoptosis, and drug resistance [115]. Their selective enrichment underscores the role of exosomes in shaping the tumor microenvironment and promoting therapy evasion. lncRNAs: Long noncoding RNAs (lncRNAs) such as H19 and MALAT1 are also preferentially sorted into exosomes, where they participate in immune escape, epithelial–mesenchymal transition (EMT), and tumor progression [116]. These lncRNAs can act as competing endogenous RNAs (ceRNAs), sponging miRNAs and thereby indirectly regulating gene expression in recipient cells. circRNAs: Circular RNAs (circRNAs) represent a class of covalently closed, highly stable RNAs that are increasingly recognized as key exosomal regulators. Within vesicles, they modulate miRNA activity by functioning as sponges or decoys, influencing gene expression networks in cancer, cardiovascular, and neurodegenerative diseases [117]. Their stability makes them promising biomarkers for diagnostic applications. DNA: Interestingly, double-stranded genomic and mitochondrial DNA have been detected in exosomes, suggesting roles in immune activation, inflammatory signaling, or potential horizontal gene transfer [118].

Tumor-derived exosomes enriched in lactate and glycolytic enzymes drive metabolic rewiring of stromal and immune cells, enhancing glycolysis and suppressing antitumor immunity [119, 120]. Amino acid cargo contributes to nutrient reallocation, while metabolic enzymes modulate systemic energy balance, contributing to cachexia in advanced malignancy [121–123]. These findings underscore metabolites as an overlooked but crucial dimension of exosomal heterogeneity. Exosomes also carry nuclear and mitochondrial DNA (mtDNA), both in linear and circular forms. Exosomal DNA can be functionally transferred to recipient cells and may be involved in inflammation, oncogenesis, or cellular reprogramming [118]. In cancer, exosomal double-stranded DNA (dsDNA) reflects the mutational landscape of the parent tumor and enables non-invasive genomic profiling [124]. This has major implications for liquid biopsy development and treatment stratification. Recent studies also suggest that exosomal DNA can serve as danger-associated molecular patterns (DAMPs), an innate immune response, activating immune receptors such as cGAS-STING pathways in macrophages, linking exosomal cargo to immune surveillance and inflammation [125], and contributing to disease pathogenesis. These findings highlight the functional versatility of exosomal nucleic acids as mediators of intercellular communication, with implications for diagnostics, disease monitoring, and therapeutic design.

Lipids

Lipidomics reveals that exosomes are enriched in sphingomyelin, ceramide, cholesterol, phosphatidylserine, and gangliosides, which contribute to their membrane stability and functional delivery [44]. The ceramide pathway is also critical in ESCRT-independent biogenesis. Exosomal lipids influence vesicle fusion, uptake, and bioactivity in target cells.

Subpopulation heterogeneity and nanoparticle classification

Emerging studies increasingly demonstrate that exosomes are not a uniform population but rather comprise multiple subpopulations that differ in size, density, surface composition, and molecular cargo. This heterogeneity reflects the complex regulatory mechanisms underlying vesicle biogenesis and their diverse biological roles. For instance, large exosomes (Exo-L) and small exosomes (Exo-S) differ not only in size but also in protein content and functional potential. Exo-Ls tend to carry higher levels of cytoskeletal and metabolic proteins, whereas Exo-Ss are enriched in signaling molecules and RNA species involved in intercellular communication. These compositional differences translate into distinct biological effects, influencing how each vesicle subtype interacts with recipient cells and contributes to physiological or pathological processes.

The advent of single-vesicle profiling technologies, such as ExoView and nano-flow cytometry, now enables the classification of exosomes based on multiplexed surface markers and cargo signatures [126]. These techniques have made it possible to dissect vesicle subpopulations with unprecedented precision, identifying unique phenotypes that were previously masked in bulk analyses. Moreover, exosome subtypes display differential targeting and biodistribution, which carry significant implications for drug delivery and diagnostic applications. Certain exosomal populations preferentially home to specific tissues or cell types, guided by their surface tetraspanins or integrin profiles, offering potential for tailored therapeutic targeting or precision liquid biopsies.

Despite these advances, exosome heterogeneity remains a major barrier to establishing consistent functional classifications. Traditional bulk analyses often obscure biologically distinct subpopulations, resulting in incomplete or overlapping definitions [127]. To overcome this limitation, emerging single-cell RNA sequencing (scRNA-seq) and spatial transcriptomics approaches are being employed separately or in combination to map exosome-producing cells and their tissue-specific secretion patterns with high resolution [128, 129]. By integrating these methods with nanoparticle tracking, proteomics, and advanced machine learning algorithms, researchers can now resolve functional subgroups with greater molecular fidelity, improving both diagnostic precision and therapeutic targeting. Together, these integrative technologies are transforming our understanding of exosomal diversity—supporting the concept of “functional heterogeneity” as a dynamic, context-dependent hallmark of extracellular vesicle biology.

Overall, the heterogeneity of exosomes reflects variations in their cellular origin and biogenetic regulation, and advances in single-vesicle profiling technologies now enable precise resolution of these distinct subpopulations for improved diagnostic and therapeutic interpretation.

Databases and bioinformatic resources

Several curated databases now catalog exosomal components from diverse organisms and cell types, providing critical tools for data mining, biomarker validation, and systems-level analysis. ExoCarta remains one of the most comprehensive and widely used databases, compiling experimentally validated data on exosomal proteins, lipids, RNAs, and metabolites [130]. It serves as a foundational reference for identifying conserved or cell-type–specific vesicle cargos, supporting hypothesis generation and candidate biomarker discovery. Vesiclepedia extends this concept as a community-annotated repository that includes not only exosomes but also microvesicles and apoptotic bodies, offering a broader landscape of extracellular vesicle content [131]. Its community-driven updates ensure that newly reported findings are rapidly incorporated, enabling more accurate cross-study comparisons. EVpedia focuses primarily on proteomics-based datasets, providing integrative tools for comparative analysis across species, experimental platforms, and disease models [132]. Its strength lies in enabling large-scale meta-analyses that reveal functional enrichment patterns and signaling networks associated with extracellular vesicles. ExoRBase is a specialized RNA-centric resource that catalogs circRNAs, miRNAs, and lncRNAs identified in human blood exosomes [133]. It facilitates the exploration of RNA-based biomarkers and regulatory interactions, particularly in cancer and cardiovascular diseases. These databases provide essential bioinformatic infrastructure that supports cross-study validation, multi-omics integration, and translational interpretation of exosomal data, accelerating the identification of clinically relevant biomarkers and therapeutic targets.

Advanced profiling technologies

Recent advances in high-throughput and nanoscale analytical technologies have revolutionized the profiling of exosomal cargo, enabling comprehensive characterization of their molecular and structural diversity (Fig. 5). These innovations allow researchers to move beyond bulk analyses toward more detailed, quantitative, and single-vesicle insights. The innovations are: Mass Spectrometry (MS): Techniques such as label-free MS and tandem MS (MS/MS) now permit deep proteomic and lipidomic profiling of exosomes, revealing thousands of components in a single run. These approaches also detect post-translational modifications (PTMs) such as phosphorylation, glycosylation, and ubiquitination, providing insight into the regulatory mechanisms that shape exosomal function. Next-Generation Sequencing (NGS): NGS has become the gold standard for identifying coding and noncoding RNAs (miRNAs, lncRNAs, circRNAs) contained within exosomes. Its high sensitivity and resolution allow quantitative detection of rare transcripts, supporting the discovery of disease-specific RNA signatures with diagnostic and prognostic value. Nanoparticle Tracking Analysis (NTA) and Flow Cytometry: These complementary tools enable quantitative assessment of vesicle concentration, size distribution, and surface marker expression. NTA provides accurate particle sizing and tracking, while high-resolution or imaging flow cytometry allows multiparametric phenotyping of distinct exosome subpopulations based on tetraspanins and other surface proteins. Microfluidic Platforms: Recent lab-on-chip microfluidic technologies facilitate single-vesicle isolation, analysis, and sorting, often integrating molecular assays such as RNA detection or immunocapture. Their minimal sample requirement and automation potential make them highly suitable for clinical translation and point-of-care diagnostics. Atomic Force Microscopy (AFM) and Cryo-Electron Microscopy (Cryo-EM): These advanced imaging modalities provide nanoscale visualization of exosomal morphology, surface topology, and cargo organization. Cryo-EM, in particular, preserves vesicle ultrastructure in near-native conditions, while AFM reveals mechanical properties relevant to uptake and membrane fusion dynamics. These multi-platform analytical tools offer an integrated understanding of exosomal composition, structure, and function, driving innovation in both biomarker discovery and therapeutic vesicle engineering.

Role of artificial intelligence and machine learning

Artificial intelligence (AI) and machine learning (ML) have become powerful tools for bioprospecting natural products [134] and in analyzing high-dimensional exosome datasets, enabling deeper insights into vesicle composition, function, and clinical relevance. By uncovering complex, nonlinear patterns within large datasets, these computational approaches enhance both the predictive and diagnostic value of exosomal biomarkers.

Artificial intelligence (AI) and machine learning (ML) approaches have accelerated exosome biomarker discovery by analyzing complex, high-dimensional data from proteomic, transcriptomic, and lipidomic studies. However, integrating multi-omics datasets introduces significant computational and biological noise, arising from variable data scales, inconsistent sample preparation, and overlapping molecular features. This heterogeneity can reduce model reliability and predictive accuracy, leading to overfitting or unstable feature selection across studies. Recent analyses demonstrate that algorithmic optimization—including dimensionality reduction, ensemble modeling, and advanced feature-selection techniques such as LASSO regression or random-forest filtering, Targeted Imaging, or gene-based algorithms like ESTIMATE—is crucial to mitigate these issues and improve reproducibility [135, 136]. Emerging frameworks that employ multi-layer neural networks and graph-based learning show promise for harmonizing diverse datasets while preserving biologically meaningful patterns. Ultimately, enhancing model transparency, reproducibility, and cross-platform validation will be essential for translating AI-driven exosome biomarker signatures into clinically actionable diagnostics.

AI/ML roles are briefly discussed: Biomarker Discovery: ML algorithms are increasingly used to identify discriminative panels of exosomal miRNAs capable of differentiating between disease states, such as early versus advanced cancer stages or malignant versus benign lesions [137]. These data-driven models improve diagnostic accuracy and may help establish standardized exosome-based liquid biopsy signatures. Classification and Clustering: Unsupervised learning algorithms, including hierarchical clustering and principal component analysis, have proven useful in revealing distinct exosome subpopulations based on proteomic and transcriptomic profiles [138]. This enables the discovery of vesicle subsets associated with specific cell types, pathophysiological conditions, or therapeutic responses. Integrative Multi-Omics: AI supports the integration of proteomic, lipidomic, and transcriptomic datasets, identifying co-regulated molecular networks and functional pathways [139]. Such integrative modeling provides a systems-level view of exosome biology, linking molecular heterogeneity to cellular function and disease mechanisms. Predictive Modeling: Deep learning frameworks, such as convolutional and recurrent neural networks, are now being applied to predict disease onset, classify disease subtypes, and monitor treatment response based on exosomal profiles [140, 141]. These models have shown promise in the early detection of neurodegenerative and oncologic disorders, highlighting AI’s potential for precision diagnostics and personalized medicine. AI and ML approaches are transforming exosome research—from descriptive molecular profiling to predictive, integrative, and clinically actionable analytics—bridging the gap between big data and translational medicine. Advances in high-throughput profiling and integrative bioinformatics have illuminated the molecular landscape of exosomes, enabling systematic exploration of their diagnostic signatures and therapeutic relevance.

High-throughput molecular profiling has unveiled the intricate composition of exosomal cargo, bridging the gap between molecular regulation and clinical application. These integrated datasets now form the foundation for biomarker validation, disease stratification, and the rational design of exosome-based therapeutics. By integrating complex proteomic, lipidomic, and transcriptomic datasets, machine learning enables more accurate discrimination of disease-specific exosomal signatures and subpopulation profiles, reinforcing its essential role in multi-omics interpretation [142].

Diagnostic roles of Exosomal bioactive molecules

Exosomes hold significant potential in diagnostics due to their stability, abundance in body fluids, and their stable encapsulation of molecular signatures, reflecting the physiological and pathological states of their cells of origin (capacity to mirror the molecular status of their parent cells). They are now being explored in liquid biopsies for cancer, neurodegenerative diseases, and cardiovascular conditions (Fig. 7). For example, GPC1 + exosomes show high specificity for pancreatic cancer, while exosomal PCA3 RNA offers improved sensitivity over PSA in prostate cancer [143, 144]. In neurodegeneration, neuron-derived exosomes carrying phosphorylated tau and α-synuclein allow earlier detection than traditional CSF assays (125, 126]. These advantages make exosomes powerful tools for minimally invasive, real-time disease monitoring (Fig. 4; Table 2).

Fig. 7.

Diagnostic roles of exosomes using a blood-based source exosome – The image is a schematic illustration summarizing the diagnostic roles of exosomal bioactive molecules in liquid biopsy applications. It presents three clinically relevant domains arranged from left to right, each represented by a simple medical icon and a labeled text box. Schematic overview of exosome-based liquid biopsy applications. Circulating exosomes in blood serve as stable carriers of tumor-derived RNAs and proteins, enabling sensitive cancer diagnostics (e.g., GPC1⁺ exosomes in pancreatic cancer and exosomal PCA3/TMPRSS2: ERG in prostate cancer). Neuron-derived exosomes containing pathological α-synuclein and phosphorylated tau support early and reliable biomarker detection in Parkinson’s disease. Exosomal proteins and miRNAs in circulation further reflect cardiovascular and metabolic risk. Collectively, the figure highlights exosomes as minimally invasive biomarkers that mirror disease-specific molecular states for real-time clinical monitoring

Table 2.

Key diagnostic performance indicators

Biomarker	Disease	AUC	Sensitivity	Specificity	Reference
Exosomal GPC1	Pancreatic cancer	0.99	100%	96%	[113]
Exosomal PCA3 + TMPRSS2: ERG RNA	Prostate cancer	0.84	81%	90%	[145]
Plasma exosomal tau	Alzheimer’s disease	0.88	85%	80%	[146]
Exosomal miR-122 and miR-192	Type 2 Diabetes/NAFLD	0.91	87%	85%	[147]
Exosomal α-synuclein	Parkinson’s disease	0.83	82%	78%	[148]

Exosome-Based liquid biopsy

Exosomes present in bodily fluids—including blood, urine, saliva, and cerebrospinal fluid—offer a non-invasive and repeatable source for molecular diagnostics (Fig. 6). Their cargo, composed of nucleic acids, proteins, and lipids, serves as a “molecular fingerprint” of diseased cells, particularly cancerous and degenerative cells [149]. Exosomes offer major advantages for liquid biopsy applications because their protected lipid bilayer preserves disease-specific biomolecules, enabling highly sensitive and specific detection of tumor-derived RNAs and proteins in accessible biofluids [150]. For example, GPC1⁺ exosomes show near-perfect specificity for pancreatic cancer, and exosomal PCA3 and TMPRSS2: ERG RNA outperforms PSA in prostate cancer detection, while circulating exosomal tau and α-synuclein enable earlier diagnosis of neurodegenerative disease compared to conventional CSF assays [145, 151]. These features underscore their strong diagnostic potential for non-invasive cancer detection and real-time disease monitoring.

Fig. 6.

Molecular Cargo Profiling - Exosomes derived from multivesicular bodies carry proteins, lipids, and nucleic acids (mRNA, miRNA, lncRNA, circRNA). These cargos are analyzed using profiling tools such as mass spectrometry, next-generation sequencing, and microfluidics, with data curated in databases like ExoCarta to support functional annotation and biomarker discovery

Oncology applications

Exosomal RNAs and proteins are increasingly recognized as valuable biomarkers for cancer detection, disease monitoring, and therapy resistance profiling (Fig. 6). Their stability in circulation and ability to reflect tumor dynamics make them powerful tools for non-invasive cancer diagnostics. Non-Small-Cell Lung Cancer (NSCLC): In NSCLC, EGFR mutations such as T790M have been successfully identified in exosomal nucleic acids using droplet digital PCR (ddPCR), allowing early detection of drug resistance before clinical progression is evident [152]. This approach provides a sensitive, minimally invasive alternative to tissue biopsy, enabling real-time tracking of molecular evolution during targeted therapy. Prostate Cancer (Exosomal PCA3 and TMPRSS2): ERG fusion transcripts demonstrate over 90% specificity for prostate malignancy detection, significantly outperforming the conventional prostate-specific antigen (PSA) test [145]. These RNA-based markers offer improved diagnostic accuracy, particularly for distinguishing clinically significant cancers from benign prostatic hyperplasia, and may reduce unnecessary biopsies. Pancreatic Cancer: The exosomal surface protein GPC1 (glypican-1) has emerged as a highly sensitive and specific biomarker capable of differentiating pancreatic cancer patients from healthy individuals with near-perfect sensitivity [113]. GPC1-positive exosomes also correlate with tumor burden and disease stage, making them promising tools for early detection and longitudinal monitoring of pancreatic malignancies. These highlight the growing translational impact of exosomal biomarkers in oncology, where they hold potential to revolutionize precision diagnostics, treatment monitoring, and personalized therapeutic decision-making.

Neurodegenerative diseases

Exosomes isolated from cerebrospinal fluid (CSF) or blood plasma have emerged as promising carriers of pathological proteins such as phosphorylated tau, α-synuclein, and amyloid-β (Aβ), which mirror neurodegenerative processes occurring within the central nervous system. Their ability to cross the blood–brain barrier and reflect ongoing neuropathology makes them valuable for developing minimally invasive biomarkers of disease progression. For Alzheimer’s Disease, exosomal tau and Aβ42 levels in CSF and plasma have been found to correlate closely with disease progression and cognitive decline [146]. These vesicle-associated proteins serve as dynamic indicators of neurofibrillary tangle formation and amyloid plaque accumulation, offering potential for early detection and longitudinal monitoring. Unlike free-circulating biomarkers, exosomal tau and Aβ42 are protected from degradation, enhancing their diagnostic stability and reproducibility. For Parkinson’s Disease, elevated exosomal α-synuclein levels in peripheral blood have been proposed as a candidate diagnostic biomarker [148]. These exosomes may originate from dopaminergic neurons and can propagate misfolded α-synuclein species between cells, potentially contributing to disease progression. Quantification of α-synuclein–enriched exosomes could therefore facilitate early diagnosis and the differentiation of Parkinson’s disease from other Parkinsonian syndromes. These underscore the potential of exosome-based liquid biopsies in neurodegenerative disorders, providing a window into brain pathology through accessible biofluids, and paving the way for earlier, less invasive diagnostic strategies.

Recent evidence suggests that the pathological spread of exosomal tau and α-synuclein is closely associated with vesicular membrane receptors such as LAMP2A, which mediate their selective loading and intercellular transfer. LAMP2A, a key component of the chaperone-mediated autophagy pathway, interacts with misfolded or aggregated proteins, promoting their association with exosomal membranes and subsequent secretion [153, 154]. Once released, these exosomes facilitate prion-like propagation of tau and α-synuclein between neurons and glial cells, amplifying neurotoxicity and accelerating disease progression [155]. The coupling between LAMP2A and pathogenic proteins underscores the dual role of exosomes as both clearance mediators and vehicles of neurodegenerative spread.

Metabolic disorders

Exosomal microRNAs such as miR-122 and miR-192 are enriched in the circulation of individuals with type 2 diabetes and non-alcoholic fatty liver disease (NAFLD), suggesting their use in early metabolic monitoring [147].

Biomarker performance metrics

Evaluating the clinical viability of exosomal biomarkers involves assessing key diagnostic performance indicators (Table 2):

Exosome-based biomarkers frequently outperform conventional diagnostics. GPC1 + exosomes identify pancreatic cancer with sensitivity > 80% and specificity approaching 100%, surpassing CA19-9 [156, 157]. Exosomal PCA3 RNA outperforms serum PSA in prostate cancer detection and risk stratification [158]. In neurodegenerative disease, exosomal α-synuclein and phosphorylated tau enable earlier detection compared to CSF protein assays, highlighting the superior sensitivity and specificity of exosome-derived biomarkers [159].

These studies collectively demonstrate that exosomal biomarkers possess diagnostic accuracies comparable to or superior to conventional biomarkers, particularly when multiplexed.

Integration with AI and Multi-omics

Artificial intelligence (AI) and machine learning (ML) approaches are emerging as powerful tools in exosome biology. By integrating multi-omics data—including RNA, lipid, protein, and metabolite cargo—ML can identify predictive biomarker signatures, stratify patients into therapeutic subgroups, and forecast treatment responses [160, 161]. These approaches significantly enhance the precision of exosome-based diagnostics compared with conventional statistical methods. Advanced machine learning algorithms have been applied to exosomal data for biomarker discovery and patient stratification: Deep learning models trained on exosomal transcriptomics have been used to classify glioblastoma vs. low-grade glioma with > 90% accuracy [162]. Multi-omics integration combining proteomic and lipidomic exosome profiles enhances diagnostic granularity and reveals disease-specific signatures that may be missed by single-modality approaches [163]. The diagnostic utility of exosomes has reached significant maturity, with numerous biomarkers nearing clinical translation. Their ability to reflect tumor heterogeneity, neurodegeneration progression, or metabolic dysregulation non-invasively makes them a cornerstone in the future of precision diagnostics. Further standardization of isolation methods and expansion of AI-integrated analytical pipelines will cement their role in routine clinical workflows.

Clinical translation and trials

Several clinical trials are evaluating exosome-based diagnostics and therapies (Table 3) (Fig. 9). For example, trials using exosomal PD-L1 as a biomarker for immunotherapy response in melanoma are underway (NCT03449792) [164, 165]. In therapeutics, MSC-derived exosomes are being tested for wound healing, graft-versus-host disease, and COVID-19 lung injury (NCT04276987, NCT04313647) [166, 167]. Cancer-directed exosome vaccines carrying tumor peptides (e.g., MAGE, HER2) are also in early-phase trials [168]. These studies highlight the translational momentum of exosomes but also underscore the need for standardized production, quality control, and safety validation before widespread clinical adoption.

Table 3.

Clinical trials and translational challenges

Condition	Exosome Type	Therapeutic Cargo	Trial Phase	Registry ID/Reference
Non-small cell lung cancer (NSCLC)	Dendritic cell exosomes	Tumor antigen peptides	Phase I/II	NCT01159288
Pancreatic cancer	MSC-derived exosomes	KRASG12D siRNA	Phase I	NCT03608631
COVID-19 ARDS	MSC-derived exosomes	Native paracrine factors	Phase I	NCT04491240

Fig. 9.

Exosome clinical translation process

Therapeutic applications of exosomes

Exosomes are increasingly recognized not only as diagnostic tools but also as potent therapeutic agents (Fig. 8). Exosomes have also emerged as delivery vehicles. These intrinsic delivery advantages enable exosomes to improve therapeutic precision and effectiveness compared with conventional nanocarriers [169]. Their natural ability to mediate intercellular communication, stability in circulation, and amenability to engineering make them highly attractive for regenerative medicine, drug delivery, and immunotherapy. Native exosomes, particularly those from mesenchymal stem cells, show regenerative and immunomodulatory effects in preclinical studies [170, 171]. Engineered exosomes can be loaded with therapeutic RNAs, drugs, or proteins and targeted to specific tissues via surface ligand modification [172, 173]. Exosome-based vaccines are being trialed in cancer to deliver tumor antigens, while plant- and bacterial-derived vesicles are being explored for oral delivery platforms [174, 175]. Despite these advances, scalability, reproducibility, and biosafety remain translational challenges.

Fig. 8.

Exosomes provide a rich source of biomarkers detectable in body fluids such as blood, urine, and saliva. Their stability, molecular heterogeneity, and cell-specific cargo make them ideal for non-invasive “liquid biopsies.” Examples include GPC1 + exosomes in pancreatic cancer, PCA3 RNA in prostate cancer, and phosphorylated tau in Alzheimer’s disease. The figure illustrates both native therapeutic functions of stem-cell exosomes (regeneration, immunomodulation) and engineered strategies for precision drug delivery. It also highlights representative exosomal biomarkers with clinical diagnostic potential

Native exosomes in disease models

Native exosomes offer intrinsic safety, biocompatibility, and immune tolerance but lack cargo precision and targeting capacity [176]. However, their implementation faces challenges, including batch reproducibility, large-scale manufacturing, and regulatory classification. Unmodified (native) exosomes, especially those derived from mesenchymal stem cells (MSCs), neural stem cells, or immune cells, have shown remarkable therapeutic potential in diverse preclinical models.

Neuroregeneration

Mesenchymal stem cell (MSC)-derived exosomes play a pivotal role in neuroregeneration by promoting neurogenesis, synaptic plasticity, and axonal growth following stroke or traumatic brain injury (Fig. 8). They act through the transfer of regulatory RNAs and proteins that activate neuronal survival pathways and suppress apoptosis. For example, exosomal miR-133b has been shown to enhance neurite outgrowth and functional recovery in post-ischemic models by targeting RhoA and other inhibitory genes involved in axonal remodeling [177]. These findings highlight the therapeutic potential of MSC-exosomes as acellular alternatives to stem cell transplantation in neurorepair. Neural stem cell (NSC)-derived exosomes also contribute to central nervous system recovery by facilitating remyelination and attenuating neuroinflammation. In experimental models of multiple sclerosis and spinal cord injury, these exosomes carry neurotrophic and anti-inflammatory molecules that promote oligodendrocyte differentiation and myelin sheath restoration, while reducing microglial activation and cytokine release [178]. These studies demonstrate that stem cell–derived exosomes serve as potent nanotherapeutics capable of orchestrating tissue repair in the injured nervous system.

Wound healing and tissue regeneration

In regenerative medicine, exosomes derived from adipose- and cardiac-origin stem cells have shown significant potential in accelerating tissue repair and functional recovery (Fig. 8). In skin repair, adipose-derived stem cell (ADSC) exosomes enhance angiogenesis, fibroblast proliferation, and collagen synthesis, thereby expediting wound closure and improving scar quality. Their molecular cargo—rich in growth factors, cytokines, and miRNAs—modulates the local microenvironment to favor tissue regeneration and reduce inflammation. In cardiac repair, exosomes from cardiac progenitor cells mitigate ischemic injury by reducing infarct size, limiting fibrosis, and improving overall cardiac function. These effects are mediated by the delivery of pro-survival, pro-angiogenic, and anti-fibrotic factors, which reprogram recipient cardiomyocytes and endothelial cells to promote myocardial recovery after infarction. Together, these examples highlight the regenerative versatility of exosomes across multiple tissue systems, underscoring their role as natural mediators of repair and regeneration.

Immune modulation

Exosomes also exert profound effects on the immune system, functioning as both activators and suppressors of immune responses depending on their cellular origin. Dendritic cell (DC)-derived exosomes are enriched with antigen–MHC complexes capable of activating T-cells and eliciting anti-tumor immune responses. This principle underlies the development of exosome-based cancer vaccines, such as Dex (dendritic cell-derived exosomes), tested in melanoma clinical trials, which demonstrated safety and the potential to stimulate cytotoxic T-cell activity against tumor antigens. Conversely, MSC-derived exosomes exhibit immunosuppressive and anti-inflammatory properties, making them beneficial in conditions such as graft-versus-host disease (GVHD) and autoimmune disorders. Although MSC-derived exosomes demonstrate strong immunosuppressive potential and have shown safety in clinical trials for graft-versus-host disease (GVHD), the risk of co-delivering pro-inflammatory cytokines such as IL-6 remains a significant concern. Recent studies propose preconditioning or purification strategies to minimize this effect—such as hypoxic pre-treatment, cytokine depletion via ultrafiltration, or affinity-based removal of inflammatory proteins before administration [179, 180]. These approaches enhance the therapeutic purity and safety of MSC exosomes by selectively preserving immunoregulatory molecules (e.g., TGF-β, IL-10) while minimizing unintended activation of inflammatory pathways. Continued optimization of exosome bioprocessing and quality control will be essential to balance efficacy with safety in future clinical applications. They deliver immunoregulatory cytokines, microRNAs, and enzymes (e.g., IDO, TGF-β) that suppress T-cell proliferation and modulate macrophage polarization toward a regulatory phenotype. Collectively, exosomes are established as versatile immunomodulatory agents capable of fine-tuning immune balance—either stimulating immune defense against tumors or promoting immune tolerance in inflammatory and autoimmune settings.

Engineered exosomes and nanovesicles

Engineered exosomes, generated via surface ligand modification or genetic programming of parental cells, enable tailored delivery of RNAs, proteins, and drugs [181]. To overcome limitations of native exosomes (e.g., low yield, non-specificity), engineered exosomes are developed for targeted therapy and enhanced cargo delivery.

Cargo loading techniques

Efficient loading of therapeutic molecules into exosomes is critical for enhancing their clinical utility as drug delivery vehicles. Several strategies have been developed to incorporate small molecules, nucleic acids, and proteins into exosomal compartments, each with unique advantages and limitations (Fig. 8). The techniques are as follows: Passive Loading: This approach relies on the simple incubation of exosomes with therapeutic agents such as drugs (e.g., paclitaxel) or nucleic acids (e.g., siRNA), allowing the cargo to diffuse across the lipid bilayer into the vesicle interior. While technically straightforward and gentle, passive loading often results in variable encapsulation efficiency, depending on cargo hydrophobicity and exosome membrane permeability. Nevertheless, it remains a practical method for small, lipophilic compounds that can readily partition into the vesicular membrane. Active Loading: To achieve higher encapsulation efficiency, active loading techniques such as electroporation and sonication are employed to temporarily disrupt the exosomal membrane, facilitating the entry of therapeutic siRNAs, CRISPR-Cas9 components, or chemotherapeutic agents. These methods offer improved delivery capacity but may affect vesicle integrity or cargo stability, necessitating optimization to maintain bioactivity and reproducibility. Producer Cell Engineering: A more biologically integrated strategy involves genetic modification of donor (producer) cells to enable the endogenous packaging of therapeutic RNA or proteins into exosomes. By engineering the cellular biosynthetic machinery, researchers can achieve highly specific, biologically compatible, and reproducible cargo loading, making this approach particularly valuable for scalable manufacturing and precision therapeutics. Overall, each loading strategy presents trade-offs between efficiency, vesicle stability, and scalability, requiring careful selection based on therapeutic context.

Surface modification and targeting

Surface engineering of exosomes is an essential step toward achieving targeted delivery and enhanced cellular uptake in therapeutic applications (Fig. 8). By modifying the exosomal membrane, researchers can control biodistribution, improve tissue specificity, and facilitate intracellular delivery of cargo. Modification strategies include: Ligand Display: One of the most widely explored strategies involves engineering exosomes to display targeting ligands—such as RVG peptides for neuronal targeting or antibody fragments for specific receptor recognition—on their surface. This modification enhances cell-specific uptake and minimizes off-target effects, making exosomes suitable for precision drug delivery to otherwise inaccessible tissues, including the central nervous system. pH-Sensitive and Chemical Conjugation Approaches: Additional strategies employ pH-sensitive fusogenic peptides or click chemistry–based surface conjugation to improve endosomal escape and targeting precision. These chemical or peptide modifications promote membrane fusion under acidic conditions or enable covalent attachment of targeting moieties, ensuring that therapeutic cargo is efficiently released at the desired intracellular site. These approaches—spanning from cargo loading to surface modification—underscore the engineering versatility of exosomes as next-generation delivery platforms, combining biological compatibility with tunable design for targeted and effective therapeutic outcomes.

Exosome-Mimetic nanovesicles

Exosome-mimetic nanovesicles can be produced by extrusion of parent cells or synthetic lipid assembly [182, 183]. These mimetics retain the targeting and uptake properties of natural exosomes while providing higher yield and better scalability. As such, they represent a promising alternative for industrial-scale therapeutic applications [183] (Fig. 8). Synthetic or semi-synthetic nanovesicles with exosome-like properties can be mass-produced. These include: Extrusion-based nanovesicles derived from cell membranes. Hybrid exosome-liposome constructs combining natural targeting with improved scalability. These synthetic systems outperform natural exosomes in production scalability and batch uniformity while maintaining biological targeting features essential for therapeutic use.

Risks associated with natural and engineered exosomes

Both natural and engineered exosomes face translational barriers. Natural exosomes are biocompatible but heterogeneous, raising concerns about batch variability, off-target effects, and inadvertent delivery of oncogenic or immunosuppressive molecules [173, 181]. Engineered exosomes offer precision cargo and targeting but face risks of immunogenicity, altered biodistribution, and manufacturing inconsistencies [184]. Upon introduction into the human system, patients may face immune rejection, rapid clearance, or unanticipated inflammatory responses [173, 185]. Experimenters, on the other hand, contend with challenges of large-scale production, reproducibility, and regulatory uncertainty [186]. Routes of administration explored in preclinical and clinical studies include intravenous infusion (systemic delivery), intranasal delivery (for CNS targeting), intratumoral injection (local effect), and oral formulations (using plant-derived vesicles). Each carries unique safety considerations related to biodistribution and clearance (Table 3).

Early clinical readouts from phase I/II trials

NSCLC – dendritic-cell exosomes (Dex) loaded with tumor antigens (NCT01159288; Phase II). Manufacturing was feasible at a multicenter scale (batch-release success ≈ 89%). Treatment was generally well tolerated (mostly grade 0–1 AEs; one grade-3 transaminitis). Although the primary endpoint (≥ 50% PFS at 4 months) was not met (observed 32%; median PFS 2.2 months; no RECIST responses), correlative immune data showed boosted NKp30-dependent NK-cell function in a subset, associated with longer PFS and with higher MHC-II/BAG6 content on Dex. Overall: safe and feasible; immunologic activity detected; limited clinical efficacy signal so far [187].

Pancreatic cancer – MSC-derived exosomes carrying KRAS^G12D siRNA (“iExosomes”; NCT03608631; Phase I). First-in-human, dose-escalation study designed to define safety and dose; as of the latest publicly available records, no peer-reviewed outcomes have been posted. Preclinical data supporting this approach are strong, but clinical safety/efficacy remain unreported [188, 189].

COVID-19 ARDS – inhaled MSC-exosomes (NCT04491240; Phase I, results posted). Registry results and narrative reviews indicate acceptable short-term safety of aerosolized MSC-exosomes with signals of improved oxygenation/clinical status in small, open-label cohorts; however, sample sizes were modest, designs were uncontrolled, and endpoints varied, so definitive efficacy cannot be inferred. Larger, randomized trials are still needed [167, 190].

Steroid-refractory graft-versus-host disease (GvHD) – MSC-exosomes (Phase II listed; mixed early clinical experience). Evidence to date includes compassionate-use cases with clinical improvement and robust preclinical efficacy in murine GvHD models; systematic reviews highlight the rationale but also the paucity of controlled human data. Phase II trials have been registered, but peer-reviewed outcome reports remain scarce, and standardized products/doses are heterogeneous [191, 192].

Safety, Feasibility, and Early Clinical Outcomes of Exosome-Based Therapies - DC-exosome vaccination in NSCLC and MSC-exosomes by inhalation/infusion show acceptable tolerability, and GMP-style manufacturing/release testing is achievable. Efficacy signals are preliminary and context-dependent (e.g., NK-cell activation with Dex in NSCLC) and have not yet translated into consistent clinical benefit in early studies. The field needs larger randomized trials with harmonized products, doses, release criteria, and endpoints to establish comparative efficacy. In parallel, transparent posting of results (including negative or neutral outcomes) will be essential to reduce publication bias and guide rational trial design [193, 194].

Translational barriers

Standardization of isolation methods: Ultracentrifugation, size exclusion chromatography, and precipitation kits vary in yield and purity. Heterogeneity and scalability: Exosome populations are heterogeneous, and scalable production without losing bioactivity remains difficult. Regulatory classification: Exosomes straddle the categories of biologics, gene therapy, and nanomedicine, creating complex regulatory hurdles. Cargo consistency: Ensuring batch-to-batch consistency of exosome cargo is critical for reproducibility and safety.

Therapeutic applications of exosomes are progressing rapidly, supported by compelling preclinical efficacy and early-phase human trials. Engineered exosomes offer new frontiers in precision medicine, from targeted gene silencing to immune activation. However, scalable production, rigorous quality control, and regulatory harmonization are essential steps for clinical translation. As technologies advance, exosomes may serve as a modular platform for next-generation biologics.

Limitations (Technical and clinical Challenges) in Exosome-Based applications

Despite their promising potential in diagnostics and therapeutics, the clinical translation of exosome technologies faces significant technical, biological, and regulatory hurdles (Fig. 10). These challenges span the entire development pipeline—from isolation and purification to scalability, cargo consistency, targeting, and regulatory acceptance.

Fig. 10.

Pipeline Schematics, technical bottlenecks, and solution - Despite their promise, exosome-based applications face technical hurdles, including low yield, isolation heterogeneity, cargo variability, off-target issues, safety/regulatory issues, quality control issues, and scalability issues. This figure outlines major bottlenecks across the pipeline—ranging from isolation and characterization to therapeutic application—and proposes solutions such as microfluidics, standardized protocols, exosome mimetics, and synthetic biology approaches

Isolation and purification

Accurate and reproducible exosome isolation is foundational to any downstream application, yet remains technically challenging. Current Techniques – (i). Differential ultracentrifugation is the gold standard, but it is time-consuming and often yields impure vesicle populations. (ii). Size exclusion chromatography (SEC) offers better purity but lower yield. (iii). Immunoaffinity capture provides specificity (e.g., CD63, CD9 markers), but scalability is limited. Microfluidic platforms and acoustic sorting are emerging as high-throughput, label-free methods. Key Challenges with isolation and purification are: (i) Co-isolation of non-exosomal vesicles (e.g., apoptotic bodies, microvesicles) blurs data interpretation. (ii). Batch-to-batch variability compromises reproducibility and clinical reliability.

Scalability and manufacturing

The low yield of exosomes per cell and the lack of standardized bioreactors hinder their large-scale production. Issues associated with it include: (i). Cell source limitations - MSCs or other primary cells can undergo senescence or phenotypic drift over passages; (ii). Production variability - Influenced by cell state, culture conditions, and environmental stress; (iii) Storage instability - exosomes can degrade or aggregate, affecting therapeutic efficacy. Likely solutions in progress include: (i). Use of cell lines engineered for high-yield exosome production; (ii) Development of hollow fiber bioreactor systems for continuous vesicle harvest; (iii) Lyophilization and cryopreservation for stable formulation.

Cargo heterogeneity and targeting specificity

Exosomes carry a broad and often unpredictable range of proteins, lipids, and RNAs. Cargo Heterogeneity: Even from the same cell line, exosome subpopulations differ in size and composition. Post-translational modifications, RNA motifs, and lipid interactions contribute to cargo variability. However, machine learning–aided omics profiling is helping to identify cargo patterns relevant to disease or therapy. Targeting Specificity: Native exosomes exhibit limited homing ability and may accumulate in non-target organs (e.g., liver, spleen). Engineered exosomes can display ligands (e.g., RGD, RVG peptides), but off-target effects and endosomal trapping remain issues.

Regulatory and immunological concerns

Exosomes occupy a regulatory gray area between biologics, cell therapy, and nanomedicine, complicating clinical approval pathways. (i). Regulatory Barriers: A key barrier to clinical translation is the lack of harmonized regulatory frameworks. Exosomes straddle definitions between biologics, drugs, and devices, complicating approval. Standardization of isolation, characterization, and quality control is urgently required to enable regulatory approval and ensure patient safety [195, 196]. No harmonized international guidelines on exosome characterization, potency assays, or release criteria. Ambiguity in whether exosomes are classified as drugs, biologics, or devices varies by jurisdiction (e.g., FDA vs. EMA). (ii). Immunogenicity & Safety Barriers: Cargo heterogeneity introduces risks of off-target effects, batch variability, and unforeseen immunogenicity. Synthetic or hybrid mimetics may pose additional safety concerns, such as horizontal gene transfer. Rigorous preclinical evaluation, safety profiling, and standardized manufacturing protocols are essential to mitigate these risks [184, 197]. Generally considered low-immunogenic, but exosomes from certain sources (e.g., tumor-derived) may carry pro-inflammatory or oncogenic factors. Biodistribution, toxicology, and long-term clearance studies are insufficiently documented.

Quality control and standardization

Reliable translation requires robust quality control (QC) and standard operating procedures (SOPs). Key quality metrics include: Particle size and concentration (NTA, TRPS), Marker expression (CD9, CD63, TSG101), Absence of contaminants (e.g., albumin, apoptotic bodies), Consistency of functional outcomes (e.g., angiogenesis, immunomodulation). Global Standardization Initiatives: The International Society for Extracellular Vesicles (ISEV) provides guidelines (MISEV2018), but clinical-grade implementation still lags.

Technical and clinical challenges continue to impede the widespread adoption of exosome-based solutions. Breakthroughs in scalable production, isolation precision, cargo engineering, and regulatory standardization are essential to fully realize the promise of exosomes in next-generation diagnostics and therapeutics. Cross-disciplinary collaborations and harmonized guidelines will be critical in moving the field toward safe, effective, and clinically approved applications.

Future perspectives

The future of exosome research lies at the crossroads of synthetic biology, bioengineering, and artificial intelligence (AI). Advances in synthetic biology now allow precise engineering of microbial systems to produce therapeutic metabolites, expanding the scope of programmable nanosystems in medicine [197]. These innovations align with exosome-based delivery strategies, enabling the creation of vesicles tailored for drug delivery, immune modulation, and regeneration. Between 2020 and 2025, perceptions of exosomes have shifted. They are no longer seen as passive carriers but as programmable nanocarriers with immense clinical and research potential (Fig. 11). Several promising directions are emerging:

Fig. 11.

Future perspectives of exosomes

Synthetic exosome mimetics

Synthetic exosome mimetics are artificially designed vesicles that aim to replicate the physicochemical and biological properties of native exosomes, while circumventing limitations such as scalability, heterogeneity, and isolation complexity. These include: Liposome-based mimetics: Engineered lipid bilayer structures encapsulating selected proteins, RNAs, or drugs, tailored for stability and delivery efficiency. Cell-derived nanovesicles (CDNs): Produced by mechanical extrusion of cells, these mimic exosomes’ surface protein profiles and show improved yield and reproducibility [198, 199]. Hybrid vesicles: Combining natural and synthetic components to retain targeting ability and improve payload control [200]. These mimetics hold promise in targeted therapy, vaccine delivery, and immune modulation, especially when coupled with surface modifications like ligands, antibodies, or peptides.

While synthetic exosome mimetics, designer vesicles, and AI-driven approaches hold promise, significant barriers remain. Immunogenicity and clearance in vivo reduce stability and therapeutic half-life [201, 202]. Manufacturing complexity, including reproducibility across batches, complicates scale-up. Ethical and regulatory uncertainties further delay translation. These challenges highlight the need for cautious optimism and rigorous validation before clinical adoption.

Programmable and designer exosomes

Advances in molecular and synthetic biology have enabled the creation of programmable or “designer” exosomes, where cells are genetically engineered to produce vesicles with customized cargo composition and surface features. These innovations mark a major step toward precision therapeutics with controllable targeting, release, and functional outcomes. Genetic programming of producer cells allows the precise incorporation of therapeutic molecules, including RNAs (e.g., miR-124 for neuroregeneration), proteins, and even CRISPR/Cas9 components, into secreted exosomes [203]. By manipulating biosynthetic pathways and cargo-sorting signals, researchers can direct specific therapeutic payloads into vesicles, enhancing both efficiency and consistency of delivery. This approach has shown promise in promoting neuronal repair, gene editing, and modulation of disease pathways at the molecular level. Ligand display systems further improve cell-type-specific targeting by engineering the exosomal membrane to express surface fusion proteins, such as LAMP2B-linked targeting peptides or antibodies [204]. These engineered motifs enable exosomes to selectively bind to receptors on desired cells or tissues—such as neurons, tumor cells, or immune cells—thereby minimizing off-target distribution and improving therapeutic precision. Responsive release systems represent the next frontier in exosome engineering, incorporating stimuli-sensitive triggers such as pH-, redox-, or enzyme-responsive elements that regulate cargo release within pathological microenvironments [205]. For instance, tumor or inflamed tissues often exhibit acidic pH or elevated enzyme activity, which can activate these modified exosomes to discharge their contents exactly where therapeutic intervention is needed.

This level of programmability and dynamic control is transforming exosomes into smart therapeutic systems—capable of multimodal delivery, on-demand release, and low immunogenicity—thereby bridging the gap between biological vesicles and engineered nanomedicine.

Exosomes in precision medicine

Exosomes are increasingly recognized as pivotal tools in precision oncology, particularly in breast and colon cancer. In breast cancer, circulating exosomal miRNAs (e.g., miR-21, miR-1246) and HER2-enriched vesicles correlate with disease progression and therapeutic resistance, providing opportunities for individualized treatment decisions [206, 207]. In colon cancer, exosomal KRAS and p53 mutations detected in plasma-derived vesicles enable real-time monitoring of tumor evolution and resistance to targeted therapies [208, 209]. Beyond diagnostics, engineered exosomes are being investigated as precision delivery vehicles for siRNAs and chemotherapeutics directly to tumor cells, minimizing systemic toxicity [210, 211]. These applications highlight the transformative potential of exosomes in tailoring cancer care.

This figure presents a roadmap of emerging strategies expected to accelerate exosome translation into clinical practice. Exosome science is rapidly evolving toward precision medicine. Advances include AI-driven biomarker discovery, synthetic biology-based exosome engineering, organoid and lab-on-chip systems for personalized therapy, and programmable designer vesicles for targeted drug delivery. Exosome-mimetic nanovesicles are artificially generated vesicles that replicate the properties of natural exosomes but offer higher yield and greater scalability. Produced by cell extrusion or synthetic lipid assembly, these mimetics overcome some production challenges of natural exosomes while maintaining targeting and uptake functions. The figure compares natural exosomes and synthetic mimetics in terms of origin, scalability, and therapeutic potential.

AI and machine learning in exosome research

Artificial Intelligence (AI) and machine learning (ML) are reshaping how exosomal data are interpreted, integrated, and applied: High-throughput cargo profiling (proteomics, transcriptomics, lipidomics) produces large, multidimensional datasets that are increasingly being analyzed with AI to identify diagnostic signatures, predict therapeutic response, and classify disease states [212]. ML models are being trained to distinguish exosome-derived biomarkers across cancer types with high accuracy (AUC > 0.9), surpassing many traditional serum-based diagnostics [213]. Network-based systems biology approaches are uncovering hidden regulatory nodes that govern cargo sorting or target interaction, informing exosome engineering strategies [7]. AI is also being used in image analysis (e.g., NTA, TEM) to improve vesicle quantification, purity assessment, and quality control [214, 215]. These computational tools are instrumental in making exosome research scalable, reproducible, and clinically actionable.

Integration with other emerging technologies

Exosomes are being positioned at the heart of emerging technologies, including Organoids and lab-on-chip systems: Emerging organoid and lab-on-chip models provide physiologically relevant, patient-specific systems for studying exosome biology [216]. These platforms enable high-throughput screening, functional testing of engineered exosomes, and personalized therapeutic design [217, 218]. Their integration will accelerate the translation of exosome research into precision medicine applications. Exosome interactions can now be modeled in physiologically relevant 3D systems to assess drug response or biomarker dynamics. Gene editing and mRNA therapies: Engineered exosomes are showing promise as non-immunogenic vehicles for CRISPR-Cas9 or therapeutic mRNA delivery. Quantum dot and nanosensor integration: Labeling exosomes with tracking agents for real-time biodistribution and target engagement visualization in vivo.

Clinical perspectives/Vaccines and therapeutics

The clinical value of cancer vaccines remains promising but modest; only a limited number have achieved regulatory approval (e.g., Sipuleucel-T for prostate cancer), with most others showing variable efficacy in late-stage trials due to tumor heterogeneity and immune escape. Risks associated with cancer vaccines include autoimmune reactions, systemic inflammation, and inconsistent patient responses. Exosomes, as emerging therapeutic vehicles, present both opportunities and risks [219]. On the one hand, their natural origin, nanoscale size, and ability to cross biological barriers make them attractive delivery platforms [220]. On the other hand, exosomes introduced as foreign bodies may trigger complement activation, off-target uptake, or immune rejection [221]. Long-term risks include transmission of unintended oncogenic cargo [222]. Mitigation strategies include engineering exosomes with defined cargo, rigorous purification, and preclinical biosafety testing.

Outlook

The integration of synthetic biology, programmable design, and AI-driven discovery is catalyzing a new era of precision exosome therapeutics. However, challenges remain, including: biosafety of synthetic or engineered vesicles, regulatory clarity for approval of bioengineered exosome platforms, and manufacturing robustness, ensuring batch consistency and scalability.

Nonetheless, these future directions promise to transform exosomes into a clinically viable platform for diagnostics, drug delivery, and regenerative medicine, bridging biology with next-generation nanomedicine.

Cautionary considerations in mitigating limitations associated with exosomes

Despite remarkable progress in understanding exosome biology and engineering, several key challenges persist that must be carefully addressed to ensure safe, effective, and reproducible clinical translation. These limitations extend across biological, technical, and regulatory domains, underscoring the need for a more harmonized global framework in exosome research and therapeutic development.

Immunogenicity

Although many preclinical and early clinical studies report that exosomes are generally well tolerated, the introduction of vesicles from non-autologous, cross-species, or engineered sources carries the potential to elicit immune or inflammatory reactions. Residual contaminants—such as proteins, nucleic acids, or process-derived impurities—can trigger unwanted immune activation or complement responses. These risks are particularly relevant for exosomes produced via genetic modification or large-scale bioprocessing, where contaminants or altered membrane antigens may persist. To mitigate this, autologous exosome sourcing, improved purification methods, and immune-evasive surface engineering have been proposed. However, comprehensive long-term immunotoxicity studies remain scarce, and understanding the immunological consequences of repeated dosing is still an unmet research need.

In vivo stability

Exosomes face rapid clearance through the mononuclear phagocyte system (MPS) and renal filtration, leading to short circulation half-lives and reduced therapeutic efficacy. Enhancing in vivo stability through PEGylation, lipid membrane modification, or encapsulation within hydrogel or nanoparticle carriers has shown promise in prolonging systemic exposure. Yet, these approaches raise concerns about toxicity, altered biodistribution, and loss of natural tropism. Achieving an optimal balance between extended circulation and biological function remains a major design challenge. The development of controlled-release delivery systems and surface modifications that preserve exosome bioactivity is a crucial next step toward improving pharmacokinetics and therapeutic durability.

Manufacturing complexity

Producing clinical-grade exosomes at scale remains one of the most significant barriers to translation. Current isolation and purification techniques—including ultracentrifugation, size-exclusion chromatography (SEC), and microfluidic separation—yield variable results in terms of purity, yield, and functionality. Scaling these methods under Good Manufacturing Practice (GMP) conditions is technically demanding and cost-intensive. Furthermore, there is a lack of universally accepted quality control (QC) metrics to evaluate batch consistency, potency, and safety. Establishing GMP-compliant pipelines, along with validated analytical assays for exosome identity, purity, and potency, is essential for regulatory approval [198; 199]. Progress in bioreactor-based production systems, standardized isolation protocols, and the use of exosome mimetics may provide scalable and reproducible alternatives, bridging the gap between laboratory discovery and industrial translation.

Ethical and regulatory barriers

Regulatory agencies worldwide continue to debate how to classify exosome-based therapeutics—whether as biologics, drugs, or advanced therapy medicinal products (ATMPs)—creating ambiguity in the approval process. This lack of regulatory consensus delays product development and complicates commercialization. Ethical issues further emerge in the use of donor-derived or genetically modified exosomes, particularly around informed consent, donor privacy, and equitable access to future therapies. Global collaboration is urgently needed to establish standardized ethical frameworks and regulatory pathways, with organizations such as the International Society for Extracellular Vesicles (ISEV) playing a central role in harmonizing practices and definitions across jurisdictions.

Reproducibility and heterogeneity

Exosomes are inherently heterogeneous, reflecting the physiological or pathological state of their parent cells. This variability leads to challenges in reproducibility, standardization, and data comparability across laboratories. Differences in isolation methods, sample sources, and analytical platforms further compound these issues, making it difficult to define universal diagnostic or therapeutic signatures. Moreover, the heterogeneous molecular cargo within exosomes complicates potency and dose-response assessments in clinical trials. Emerging solutions include single-vesicle profiling technologies, multi-omics integration, and AI-driven classification tools, which can help identify disease-specific exosome subpopulations and reduce analytical noise. Nevertheless, reproducibility remains a cornerstone challenge for both academic and industrial stakeholders.

Conclusion

Exosomes are pivotal mediators of intercellular communication. They regulate homeostasis, drive disease processes, and offer therapeutic potential. Their cargo—proteins, lipids, RNAs, and DNAs—reflects tightly controlled biogenesis and sorting pathways. Both ESCRT-dependent and independent mechanisms ensure their production across diverse cellular states. Clinically, exosomes are being developed as non-invasive liquid biopsy tools for cancer, neurodegeneration, and metabolic diseases. Their stability, targeting ability, and natural compatibility also make them promising drug delivery vehicles for conditions such as glioblastoma, chronic wounds, and inflammatory disorders (Fig. 12). Key challenges remain: isolation methods lack standardization, cargo heterogeneity complicates analysis, and large-scale GMP production is difficult. Regulatory frameworks are still evolving. Yet advances in high-throughput profiling, machine learning, and synthetic vesicle engineering are addressing these gaps. Future progress will depend on integrating exosome biology with systems biology, nanotechnology, and precision medicine. With these innovations, exosomes are poised to transform diagnostics, enable precision therapeutics, and expand the boundaries of cell-free medicine.

Fig. 12.

Graphical Summary

Exosomes are nanoscale vesicles regulated by ESCRT pathways, RNA-binding proteins, lipid dynamics, and PTMs. They carry proteins, lipids, and nucleic acids that can be profiled for biomarker discovery and integrated into databases such as ExoCarta. These features underpin their emerging roles as liquid biopsy tools and therapeutic platforms, from regenerative medicine to engineered vesicle-based drug delivery.

Abbreviations

EVs

Extracellular Vesicles

MVBs

Multivesicular Bodies

ILVs

Intraluminal Vesicles

ESCRT

Endosomal Sorting Complex Required for Transport

GAS5

Growth Arrest–Specific Transcript 5

PD-L1

Programmed Death-Ligand 1

miRNAs

MicroRNAs

lncRNAs

Long non-coding RNAs

circRNAs

Circular RNAs

LAMP2A

Lysosome-Associated Membrane Protein 2 A

MHC

Major Histocompatibility Complex

MISEV

Minimal Information for Studies of Extracellular Vesicles

MMP-9

Matrix Metalloproteinase-9

mRNA

Messenger RNA

MSC

Mesenchymal Stem Cell

NK

Natural Killer

nSMase2

Neutral Sphingomyelinase 2

oncomiRs

oncogenic miRNAs

PTM

Post-Translational Modification

piRNAs

PIWI-interacting RNAs

QC

Quality Control

Rab

Ras-Associated Binding Protein

RBP

RNA-Binding Protein

SEC

Size-Exclusion Chromatography

STING

Stimulator of Interferon Genes

TME

Tumor Microenvironment

TSG101

Tumor Susceptibility Gene 101

VEGF

Vascular Endothelial Growth Factor

VPS4

Vacuolar Protein Sorting-Associated Protein 4

YBX-1

Y-Box Binding Protein 1

AI

Artificial Intelligence

ALIX

ALG-2–Interacting Protein X

APC

Antigen-Presenting Cell

CD

Cluster of Differentiation

circRNA

Circular RNA

cGAS

Cyclic GMP–AMP Synthase

DC

Dendritic Cell

Dex

Dendritic Cell–Derived Exosome

DNA

Deoxyribonucleic Acid

EMT

Epithelial–Mesenchymal Transition

Exo-L

Large Exosome

Exo-S

Small Exosome

GMP

Good Manufacturing Practice

HIF-1α

Hypoxia-Inducible Factor 1 Alpha hnRNPA2B1Heterogeneous Nuclear Ribonucleoprotein A2/B1

HSP

Heat Shock Protein

ISEV

International Society for Extracellular Vesicles

Author contribution

V.U.C. conceived, initiated, and developed the manuscript till completion.

Funding

This study was conducted without any specific funding from external agencies in the public, commercial, or not-for-profit sectors. The author carried out the research as part of their academic and professional responsibilities, utilizing institutional resources.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.

Ethics

The study employed a narrative review methodology that utilized publicly available data and peer-reviewed literature. Since no human participants, animals, or private data were directly involved in this research, ethical approval was not required. However, the study adhered to ethical principles for conducting and reporting research, including transparency, accuracy, and integrity in data analysis and interpretation.

Conflict of interest

The author declares no conflict of interest in relation to this study. No financial or personal relationships influenced the research process, results, or interpretation. All findings and conclusions are presented impartially.