In the 1980s, scientific exploration of extracellular vesicles (EVs) began, initially focusing on reticulocyte maturation, which revealed EVs containing transferrin receptors were crucial for modulating reticulocyte maturation [1]. Subsequently, these EVs were identified as “exosomes”, originating from fusion of multivesicular bodies (MVBs), displaying phospholipid-bilayer-enclosed spherical structures and carrying diverse cargos [2, 3].

Current understanding of EVs encompasses various subtypes released into body fluids like blood, urine, and cerebrospinal fluid, serving as rich reservoirs. These subtypes include exosomes (40–100 nm), microvesicles (MVs, 100–1000 nm), apoptotic bodies (ApoBDs, 50–500 nm), and large oncosomes (1–10 μm) [4]. These particles mediate cell-to-cell communication, delivering functional biomolecules critical for physiological functions and influencing recipient cells’ composition and function [5]. They also serve as markers for diagnosis, disease progression, and therapeutic targeting, especially in liver diseases [6, 7].

Analyzing EVs poses challenges due to their size and heterogeneity [8, 9]. Specific methodologies and functional/physical analyses like electron microscopy (EM) and nanoparticle tracking analysis (NTA) are recommended. Size exclusion chromatography and ultracentrifugation remain vital tools in EV research.

Recent research on EVs has uncovered the exploitation of cellular EV pathways by viruses, including Herpesviruses [10], as well as hepatotropic viruses [11,12,13], for replication and transmission. However, a gap exists in comprehensive reviews summarizing these latest findings and outlining future perspectives. This review aims to consolidate insights into EV pathways used by hepatotropic viruses, prospecting future research.

EVs are classified based on distinct biogenesis pathways [4] (Fig. 1). Four main types, including exosomes, MVs, large oncosomes, and ApoBDs, are identified. Among these, MVs and large oncosomes belong to the category of ectosomes. The formation of exosomes and ectosomes relies on local microdomains assembled in endocytic membranes for exosomes and in the plasma membrane for ectosomes [14]. These microdomains govern the accumulation of proteins and RNAs associated with their cytosolic surface, leading to membrane budding inward for exosome precursors, known as intraluminal vesicles (ILVs), and outward for ectosomes. These two types of vesicles differ in size, with exosomes typically ranging from 40 to 100 nm and ectosomes ranging from 100 to 10,000 nm, as well as in the mechanisms of assembly, composition, and regulation of release, although there are some partially overlapped mechanisms [14]. Further details of these vesicles are described in the section below.

Biogenesis pathways of main extracellular vesicles. The four primary types of extracellular vesicles, exosomes (40–100 nm), microvesicles (100–1000 nm), large oncosomes (1–10 μm), and apoptotic bodies (50–500 nm), are illustrated. The initiation step of exosome formation involves the creation of early endosomes, which happens subsequent to the endocytosis or uptake of extracellular fluids, particles, and viruses through receptor- and Clathrin-dependent or independent routes. Notably, if there are any plasma membrane receptors or membrane-integrated proteins located within the region of the endocytic membrane, their orientation changes from facing the outside of the cells to facing the endosomal lumen after endocytosis-mediated internalization. These receptors can then either be recycled to the plasma membrane or retained within the endosome. Exosomes are then formed through endosomal membrane inward budding to generate intraluminal vesicles (ILVs) via endosomal sorting complexes required for transport (ESCRT) complexes or an ESCRT-independent route through lipid rafts, such as the membrane-associated neutral sphingomyelinase (nSMase) and the ceramide-triggered pathway. As ILVs form, the orientation of plasma membrane receptors or membrane-integrated proteins within the endosome undergoes another change, transitioning from facing the endosomal lumen to facing the outside of the ILV. Both ESCRT-dependent and nSMase/ceramide-triggered exosomes can be inhibited by blockers such as Manumycin A, GW4869, and Altenusin. As ILVs accumulate within a single endosome during endosome maturation, the early endosome progresses into the late endosome, also referred to as the multivesicular bodies (MVBs). MVBs can proceed through two pathways: fusion with lysosomes for degradation, which involves the small GTPase Rab7, or secretion into the extracellular space as exosomes after MVB-plasma membrane fusion, a process regulated by small GTPases such as Rab11, Rab27 and Rab35. Microvesicles and large oncosomes are categorized as ectosomes, originating as outward buds from the plasma membrane. Apoptotic bodies result from the orderly fragmentation of apoptotic cells, and the formation of apoptotic bodies involves key roles played by caspase-3 substrates, including ROCK1, PANX1, and PLEXB2

Exosomes, the first discovered EV subtype [1], undergo a series of steps during biogenesis. The initiation step involves the formation of early endosomes, which occurs after the endocytosis or uptake of extracellular fluids, particles, and viruses through receptor- and Clathrin-dependent or independent routes [15]. Importantly, if there are any plasma membrane receptors or membrane-integrated proteins located within the region of the endocytic membrane, their orientation changes from facing the outside of the cells to facing the endosomal lumen after endocytosis-mediated internalization (Fig. 1). These receptors can then either be recycled to the plasma membrane or retained within the endosome. Subsequently, ILVs are formed through inward budding of the endosomal membrane [16]. As ILVs form, the orientation of plasma membrane receptors or membrane-integrated proteins within the endosome undergoes another change, transitioning from facing the endosomal lumen to facing the outside of the ILV (Fig. 1).

The process of ILV formation is facilitated by endosomal sorting complexes required for transport (ESCRT) complexes (ESCRT-0, ESCRT-I, ESCRT-II, and ESCRT-III) and the ALG-interacting protein X (ALIX)-Syntenin complex (ESCRT-dependent pathway) [16]. In this pathway, ESCRT-0, along with disassembly and deubiquitinating enzymes, as well as the ESCRT accessory protein Vacuolar Protein Sorting 4 (VPS4), cluster on the endosomal membrane at the cytoplasmic side to sort cargos [17]. The Hepatocyte Growth Factor-Regulated Tyrosine Kinase Substrate (HRS) subunit of ESCRT-0 coordinates early steps in ILV biogenesis by binding to cargoes and recruiting clathrin to the early endosome [18]. Subsequently, ESCRT-I, ESCRT-II, and ESCRT-III are sequentially recruited to maturing endosomes [17]. ESCRT-II induces the formation of ESCRT-III filaments, which sever the nascent ILVs from the endosome membrane [19]. Alternatively, the ESCRT-III complex can be recruited by ALIX, binding to lysobisphosphatidic acid on the endosomal membrane [20]. ESCRT-III may sense negative membrane curvature or promote membrane bending to drive fission [21, 22]. Notably, studies have shown that silencing ESCRT-0 proteins HRS or Signal Transducing Adaptor Molecule 1 (STAM1), or ESCRT-I subunit Tumor Susceptibility 101 (TSG101), reduces small EV secretion, suggesting redundancy in ESCRT-II and ESCRT-III components [23]. Variations on the ESCRT pathway, such as the syndecan–syntenin–ALIX pathway, are specifically compromised by knockdown of genes encoding ESCRT-I protein TSG101, ESCRT-II subunit Vacuolar Protein Sorting 22 (VPS22), or ESCRT-III filament protein Charged Multivesicular Body Protein 4A (CHMP4) [24]. Syndecan-1 interacts with syntenin and ALIX on endosomes, thereby facilitating cargo sorting [24,25,26]. Thus, syndecan-1, syntenin, and ALIX play roles in cargo sorting during this process.

Currently, it is recognized that ILV production can also occur via ESCRT-independent processes involving lipid rafts [27, 28]. These lipid rafts, rich in cholesterol and sphingolipids, are subject to the activity of the neutral sphingomyelinase (nSMase) family, membrane-bound enzymes that convert sphingolipids, specifically sphingomyelin, to ceramide, a cone-shaped rigid lipid [29, 30]. Once converted, ceramide can form lipid raft microdomains, such as ceramide-enriched membrane domains, and initiate the formation and inward budding of ILVs into the endosome. Due to their unique cone-shaped structure, ceramides induce spontaneous membrane invagination, facilitating ILV formation within endosomes and maintaining vesicle shape and structure [29,30,31].

Both ESCRT-dependent and nSMase/ceramide-triggered exosomes can be inhibited by blockers such as Manumycin A, GW4869 and Altenusin [32]. Manumycin A functions as an inhibitor for both ESCRT and nSMase/ceramide-dependent pathways [33, 34]. In the ESCRT-dependent pathway, Manumycin A targets ESCRT-0 proteins HRS, ALIX, and Rab27a, leading to the inhibition of exosome biogenesis and secretion. This inhibition primarily occurs through targeted suppression of the Ras/Raf/ERK1/2 signaling pathway [34]. Additionally, Manumycin A irreversibly inhibits nSMase, further reducing exosome biogenesis and secretion [33]. GW4869 selectively inhibits nSMase2, while Altenusin selectively inhibits a broad range of nSMases [35, 36]. These inhibitors serve as valuable tools for investigating the physiological and pathological roles of exosomes.

As ILVs accumulate within a single endosome during endosome maturation, the early endosome progresses into the late endosome, also referred to as the MVB [37, 38]. MVBs can proceed through two pathways: fusion with lysosomes for degradation, which involves the small GTPase Rab7 [39], or secretion into the extracellular space as exosomes following fusion of the MVB with the plasma membrane, a process regulated by small GTPases such as Rab11, Rab27 and Rab35 [37, 38].

MVs and large oncosomes fall under ectosomes, distinct from exosomes, originating as outward buds from the plasma membrane. Small ectosomes like MVs share machinery with exosomes, involving tetraspanin proteins, such as CD9, CD63, and CD81, which interacts with Ezrin, Radixin, Moesin (ERM) and Glu-Trp-Ile EWI Motif-Containing Protein (EWI) proteins connecting to the actin cytoskeleton and impact plasma membrane organization, signaling, cargo sorting, and vesicle budding [40]. For example, CD82 recruits the ERM protein ezrin to membrane blebs for release in ectosomes [41,42,43]. Protrusions like filopodia, cilia, and microvilli promote MV shedding after the formation of ectosomal blebs [44,45,46], e.g., Human immunodeficiency virus 1 (HIV-1) particles assembling at filopodia tips, suggesting potential contributions to both retrovirion and ectosome biogenesis [47].

Large ectosome formation, on the other hand, is less understood than exosomes and small ectosomes. It remains uncertain whether early ESCRT machinery or features of exosome or small ectosome biogenesis are involved. Actin cytoskeleton rearrangements underlie plasma membrane blebbing and scission to release large EVs, with molecular reorganizations and alterations in proteins, lipids, and electrolyte levels implicated in the process [48]. Local disassembly of the cortical actin cytoskeleton, combined with actomyosin contractility, can promote plasma membrane blebbing and the subsequent formation of large ectosomes, especially in non-apoptotic cancer cells transitioning to a more migratory and metastatic phenotype [49].

Both large and small EVs contain the lipid raft marker caveolin-1 (CAV1) [50], suggesting an association with or derivation from lipid raft-associated membrane domains. CAV1, known for regulating small EV biogenesis through cholesterol binding [51], may apply to ectosomes. However, systematic studies on cholesterol and lipid rafts in EV biogenesis are lacking, leaving the possibility that the ectosomal membrane is derived from the plasma membrane and associated lipid rafts.

The biogenesis of ApoBDs differs significantly from other EV subtypes, arising from the ordered fragmentation of apoptotic cells during programmed cell death. This process progresses through several stages, including nuclear chromatin condensation, nuclear splitting, micronuclei appearance, membrane blebbing, and cellular content splitting into ApoBDs [52, 53]. Caspase-3 substrates, including Rho Associated Coiled-Coil Containing Protein Kinase 1 (ROCK1), Pannexin 1 (PANX1), and Plexin B2 (PLEXB2), play key roles in the formation of ApoBDs [52, 53]. ApoBDs can also form from protrusions known as apoptopodia [52, 53]. For instance, ROCK1 activates actomyosin contractility, leading to blebbing either directly from the plasma membrane or from the tips of surface protrusions called apoptopodia [52, 53]. Existing data suggest that membrane blebbing is, at least in part, mediated by actin-myosin interaction [54, 55]. In normal development, most ApoBDs are phagocytosed by macrophages and cleared locally [56]. However, it has been reported that the process of apoptotic cell disassembly and the removal of apoptotic material by phagocytes are rapid, limiting the presence of ApoBDs in vivo [57, 58].

A multitude of hepatotropic viruses has been identified and extensively scrutinized for their role in precipitating liver diseases, encompassing both acute and chronic hepatitis, as well as hepatocellular carcinoma (HCC) [59]. Currently, hepatotropic virus infections persist as a formidable public health challenge, contributing significantly to morbidity and an annual global mortality of approximately 1.5 million deaths [60]. Five prevalent viruses, namely hepatitis A (HAV), hepatitis B (HBV), hepatitis C (HCV), hepatitis D (HDV), and hepatitis E (HEV), emerge as major contributors to various liver diseases, each displaying distinct geographical localizations [61]. Notably, HBV and HCV impose the most substantial socioeconomic burdens, particularly in developing regions like Africa and Asia [61]. These viruses adeptly co-opt host materials for replication, establishing prolonged persistence through varying strategies. Their life cycle commences with attachment and entry into hepatocytes, relying on unique cell surface receptors, such as sodium taurocholate cotransporting polypeptide (NTCP) for HBV and CD81 for HCV [62, 63]. Subsequently, hepatotropic viruses manipulate host transcriptional machinery and cellular resources for replication, leading to uncontrolled viral proliferation, massive hepatocyte necrosis, inflammatory infiltration, and the onset of severe conditions like cirrhosis, HCC, or other critical illnesses [59].

Growing evidence indicates that EVs can function as carriers for these viruses, directly contributing to viral replication, transmission, or pathogenesis [64, 65]. The involvement of ESCRT components in viral capsid packaging and the maturation of enveloped viruses represents the initial evidence supporting this idea [66]. Recent studies have further illuminated the presence of viral components within EVs, particularly in exosomes and MVs, revealing that hepatotropic viruses exploit EVs to replicate, transmit their genome, and establish persistent infections. These viruses can conceal within EVs through ESCRT-dependent or independent viral budding modes, evading immune detection [67, 68].

HAV, a small RNA virus classified within the Picornaviridae family, exhibits typical features characteristic of a classic non-enveloped virus. With a 7500-nucleotide positive-strand RNA genome and a diameter ranging from 27 to 32 nm, it encodes a sizable polyprotein. Upon binding to receptors such as HAV cellular receptor 1 (HAVCR1) with assistance from integrin β1, the virion enters cells [69]. Translation is initiated upon uncoating and is regulated by an internal ribosome entrance site (IRES) in the cytoplasm [70] (Fig. 2A). For translation initiation, HAV relies on intact eukaryotic initiation factor 4G (eIF4G) [71]. The translated polyprotein then undergoes proteolysis, resulting in four capsid proteins (VP1, VP2, VP3, VP4) and seven nonstructural polypeptides (2A, 2B, 2C, 3A, 3B, 3C, 3D) [72]. Notably, 2A is subsequently considered a misidentification and is re-named as pX [70].

The crosstalk of canonical and extracellular vesicle-mediated HAV life cycle. A The schematic illustrates the genome structure of HAV, consisting of four capsid proteins (VP1, VP2, VP3, VP4) and seven nonstructural polypeptides (2A, 2B, 2C, 3A, 3B, 3C, 3D). Note that 2A is functionally equivalent to pX. B Non-enveloped HAV (neHAV) enters host cells by binding to receptors such as HAV cellular receptor 1 (HAVCR1) and Integrin β1, followed by endocytosis. Although the mechanisms and players involved in this process are currently unknown, upon uncoating, HAV RNA is released into the cytoplasm and utilizes host translation machinery to produce polyprotein. After proteolytic processing, the product of 3D (RNA-dependent RNA polymerase) is employed to replicate the HAV genome, which is then assembled into pre-assembled capsids and released. The mechanisms of neHAV release remain unclear. For quasi-enveloped exosomal HAV (eHAV), it is currently considered that following HAVCR1 binding to phosphatidylserine, Clathrin/dynamin-dependent endocytosis, facilitated by integrin β1, coordinates the entry of eHAV into cells and forms endosome, regulated by Rab5C. The endocytic endosome can then fuse with other endosomes or undergo maturation to form multivesicular bodies (MVBs), which can further fuse with lysosomes, facilitated by Rab7A, to release the eHAV genome and facilitate replication, or fuse with the plasma membrane to release eHAV as exosomes, although the Rabs participating in this process remain unknown. The current understanding of eHAV biogenesis from packaged neHAV involves host proteins associated with the ESCRT, specifically VPS4B and ALIX. The viral protein pX, present on the surface of eHAV but absent in non-enveloped virions, plays a crucial role in its biogenesis, potentially through interaction with ALIX via the C-terminal portion of pX, facilitating the conversion of neHAV into intraluminal vesicles (ILVs) within an MVB. Subsequently, these ILVs are secreted as eHAV after fusion of the MVB with the plasma membrane. The question mark denotes an unknown or unclear process and molecular mechanism

Genome replication follows the standard positive-stranded RNA virus model. Capsid assembly, guided by VP0 (containing VP4 and VP2), VP3, and VP1pX, involves cleavage of VP1pX by a cellular protease, leading to the removal of pX from HAV particles [73]. Despite the availability of vaccines [