EVs are widely distributed and derived from diverse cellular sources [11]. Their biogenesis pathways exhibit remarkable heterogeneity and specificity, with distinct EV subtypes displaying unique biological characteristics attributable to their formation mechanisms.

Exosomes biogenesis initiates through the endosomal pathway, where plasma membrane invagination forms early endosomes that subsequently mature into late endosomes [12, 13]. The limiting membrane of late endosomes undergoes inward budding to encapsulate cytoplasmic components, generating intraluminal vesicles (ILVs) within multivesicular bodies (MVBs) [3]. In contrast, MVs are generated via direct plasma membrane outward budding, which packages cytosolic biomolecules before detaching from the parental cell membrane. MVs typically range from 100 nm to 1 µm in diameter [14]. Notably, exosomes and MVs share overlapping molecular compositions, including adhesion molecules, membrane receptors, tissue factors, cytoskeletal proteins, chemokines, enzymes, cytokines, and nucleic acids [15]. Apoptotic bodies, a distinct EV subtype, are produced exclusively during programmed cell death or apoptosis. With diameters of 1–5 µm, they are significantly larger than other EVs constitutively released by viable cells [14, 16].

The core molecular composition of EVs is complex and highly diversified, mainly including membrane proteins, nucleic acids, functional proteins, lipids and metabolites, etc. These molecules jointly endow EVs with biological functions and play a key role in intercellular communication [17, 18, 19, 20]. Membrane proteins are not only the surface markers of EVs such as Cluster of Differentiation 9 (CD9), Cluster of Differentiation 63 (CD63) and Cluster of Differentiation 81 (CD81), but also actively participate in the generation of EVs, screening of contents, targeted recognition and cellular uptake [21]. EVs carry nucleic acids including miRNAs, mRNA, circRNAs, and mitochondrial Deoxyribonucleic Acid (DNA). These nucleic acid molecules can be transferred to recipient cells to regulate their gene expression and functional states [22]. Additionally, circRNAs, due to their circular structure that resists degradation, exhibit more enduring effects in intercellular regulation [23]. Functional proteins such as heat shock proteins (e.g., Heat Shock Cognate 70 [Hsc70], Heat Shock Protein 70 [HSP70]) and endosomal sorting complex required for transport (ESCRT)-related proteins such as tumor susceptibility gene 101 are not only involved in the formation of MVBs, intracellular vesicles and exosomes, but also may affect the stress response and immune function of receptor cells [24]. The composition of lipids (such as cholesterol, sphingolipid and phosphatidic acid) not only maintains the structural integrity of EVs, but also participates in signal transduction and membrane fusion processes, which affect the interaction ability between EVs and target cells [25]. Recent studies have shown that EVs also carry a range of metabolites such as fatty acids, amino acids and ketones, which can reflect the metabolic status of cells and may mediate distal effects such as metabolic reprogramming, with potential pathological significance in metabolic stress diseases such as heart failure [26].

The functional implementation of EVs relies on a cascade of regulatory processes from biogenesis to targeted action. During the intracellular phase, the biosynthesis of ILVs is governed by sophisticated mechanisms, primarily involved both ESCRT-dependent and ESCRT-independent pathways. The ESCRT-dependent pathway comprises four distinct complexes (ESCRT-0, -I, -II, and -III) and accessory proteins (e.g., Vacuolar Protein Sorting 4 [Vps4], ALG-2 Interacting Protein X [ALIX]), each performing specialized regulatory functions [27, 28]. In contrast, the ESCRT-independent pathway employs alternative mechanisms such as the Syndecan-Syntenin-ALIX axis, lipid rafts, tetraspanins, and Rab family guanosine triphosphatases (GTPases), collectively contributing to ILVs formation and cargo sorting, thereby demonstrating the complexity of exosome biogenesis [29]. Microvesicle generation similarly involves coordinated action of multiple molecules. Luminal lipidated proteins (e.g., myristoylated and palmitoylated proteins) may facilitate membrane curvature, while certain ESCRT subunits (I/II/III) participate in microvesicle assembly and budding. Additionally, ceramide, which is produced through acid sphingomyelinase activation contributes to plasma membrane-derived release [29]. Apoptotic bodies emerge during late-stage programmed cell death, regulated by caspase-mediated cleavage and subsequent Rho-associated protein kinase activation [16, 30]. These factors coordinate cytoskeletal reorganization and membrane dynamics to control apoptotic body formation. Upon entering circulation, EVs undergo microenvironmental modifications including membrane’s asymmetry establishment, polarization within pH gradients, localized deformation and migration, as well as surface corona formation. When reaching target cells, integrin family proteins mediate tissue-specific homing, followed by functional cargo delivery via membrane fusion or endocytic pathways, ultimately reprogramming recipient cell signaling networks [31]. This series of precisely regulated processes not only ensures EVs’ accurate participation in intercellular communication but also provides mechanistic foundations for developing EV-based therapeutic interventions.

The extracellular matrix (ECM) plays an indispensable role in the generation of EVs [32]. Through mechanical transduction signals, the stiffness and viscoelastic properties of the ECM remodel actin cytoskeletal assembly and regulate membrane raft dynamics, thereby directly influencing vesicle formation efficiency [33, 34]. Additionally, EVs themselves participate in multiple ECM regulatory functions including matrix degradation, protein cross-linking, and calcification [35]. Notably, the ECM also modulates vesicle cargo loading by balancing non-coding RNA expression within cells [36]. For instance, after myocardial infarction, specific miRNAs carried by exosomes released by endocardial and epicardial cells require ECM-mediated intercellular communication to achieve targeted delivery [34].

CM-EVs exhibit dual regulatory roles in HF through their cargo of diverse miRNAs, proteins, and cytokines. In their pathological capacity, CM-EVs from ischemic myocardium are internalized by infiltrating monocytes, triggering increased production of pro-inflammatory cytokines and chemokines that exacerbate HF. Xia and Wang [37] demonstrated that stress-induced CM-EVs enriched with miR-21 promote post-ischemic cardiac fibrosis via the activation of protein kinase B (AKT) signaling pathway. Moreover, EVs from necrotic cardiomyocytes enhance inflammatory responses by stimulating interleukin-6 (IL-6), CC motif chemokine ligand 2 (CCL2), and CC motif chemokine ligand 7 (CCL7) release upon monocyte uptake [38]. Tang et al. [39] revealed that Peli1-induced CM-EVs carrying miR-494-3p activate cardiac fibroblasts through phosphatase and tensin homolog (PTEN) suppression and subsequent hyperphosphorylation of Akt, Mothers Against Decapentaplegic Homolog 2/3 (SMAD2/3), and Extracellular Signal-regulated Kinase (ERK), thereby accelerating fibrosis. Additional studies showed that hypertrophic CM-EVs transfer miR-155 to macrophages, inducing pro-inflammatory cytokine (IL-6/IL-8) release [40]. Conversely, cardioprotective CM-EVs deliver beneficial effects: miR-30d-containing vesicles attenuate cardiac fibrosis, apoptosis, and hypertrophy to improve the function [41]. Senesi et al. [42] reported that induced CM-EVs carrying miR-24-3p suppress cardiac fibrosis, mitigating HF progression. Healthy adult CM-EVs multi-targetedly inhibit pro-fibrotic pathways Mitogen-activated Protein Kinase (MAPK), mammalian Target of Rapamycin (mTOR), Janus kinase/signal transducer and Activator of Transcription (JAK/STAT), Transforming Growth Factor Beta (TGF$\beta$), Phosphatidylinositol 3-kinase (PI3K)/Akt, reversing fibroblast activation, reducing extracellular matrix deposition, and promoting angiogenesis [23]. Zhang et al. [43] identified let-7b-5p as mediating cardiac remodeling through small EV-mediated delivery. Emerging evidence showed that CM-exosomal miR-22-3p targets acyl-coA synthetase long-chain family member 4 to inhibit ferroptosis susceptibility in ischemic HF [31]. These findings underscore CM-EVs’ functional dichotomy in HF pathogenesis and therapy. Their cargo-specific heterogeneity and molecular mechanisms warrant further exploration to develop precision interventions. Current evidence showed CM-EVs as both disease mediators and therapeutic vectors, with their net effects depending on parental cell status and pathological context.

Similar to CM-EVs, CF-EVs exhibit dual functional roles in HF. CF-EVs contribute to HF progression by delivering pathological cargo to cardiomyocytes. Qiao et al. [44] demonstrated that CF-EVs transport miRNAs that activate hypertrophic signaling pathways, promoting cardiomyocyte hypertrophy. Under oxidative stress, miR-27a enriched fibroblast-derived exosomes suppress PDZ and LIM domain 5 (PDLIM5) expression, exacerbating cardiac hypertrophy in a myocardial infarction induced chronic HF model [45]. In patients with HF post-acute myocardial infarction, circulating exosomal levels of miR-192, miR-194, and miR-34a are significantly elevated, with miR-194 and miR-34a correlating positively with left ventricular diastolic dimensions and negative with ejection fraction [46]. Additionally, Tian et al. [47] reported that stress-responsive miRNAs (e.g., miR-28a, miR-34a) are upregulated in cardiac fibroblasts, packaged into EVs, and secreted to suppress nuclear factor erythroid 2-related factor 2 (Nrf2) translation, directly impairing cardiomyocyte function. Conversely, EVs may exert beneficial effects under certain conditions. Studies have shown that exosomes derived from induced pluripotent stem cells differentiated cardiac fibroblasts from HF patients exhibit reduced miR-22 but elevated miR-371/372/373 expression, suggesting a role in promoting myocardial repair [48]. Bang et al. [49] identified that CF-EVs carrying miR-21-3p mediate cardiomyocyte hypertrophy, yet this mechanism may also be harnessed for therapeutic regeneration. These findings highlight CF-EVs as potential targets for modulating cardiac remodeling and fibrosis in HF.

EC-EVs, as a key cellular component maintaining cardiovascular homeostasis, demonstrate dual regulatory functions in the pathogenesis of HF, working synergistically with cardiomyocyte- and cardiac fibroblast-derived EVs to participate in the complex pathological network of HF. Suades et al. [50] found that circulating extracellular vesicles (cEVs) derived from endothelial cells are mainly phosphatidylserine circulating extracellular vesicles (PS^–-cEVs) and significantly elevated in chronic HF patients, though their precise mechanistic role in HF remains unclear. Regarding HF progression promotion, studies have shown that high glucose-induced CaMK2a/O-GlcNAcylation positive feedback loop in endothelial cells leads to sustained calcium/calmodulin-dependent protein kinase II$\alpha$ (CaMK2a) activation, subsequently generating plasma small EVs (sEVs) [51]. Arterial endothelial cells serve as the primary source of miR-15-16 in these sEVs, which exert persistent detrimental effects on cardiomyocytes and induce cardiac dysfunction in healthy animals. Conversely, in HF antagonism, research demonstrates that glucose-deprived H9C2 rat cardiomyocytes can modulate endothelial cell-released exosomes overexpressing various miRNAs (miR-17, miR-19a, etc.), thereby promoting angiogenesis [52]. Furthermore, under physiological conditions, EC-EVs influenced by cardiomyocytes exhibit a high proportion of vesicular transport-related proteins. Van Balkom et al. [53] discovered that endothelial cell EC-derived exosomes enriched with miR-214 promote angiogenesis and prevent cell cycle arrest in recipient ECs, potentially exerting protective effects by improving myocardial blood supply. Another study revealed that circulating EC-EVs from hypertensive individuals may potentially increase HF risk by upregulating proteins associated with cardiac hypertrophy and fibrosis while downregulating nitric oxide synthase (NOS) expression [54]. Comprehensive elucidation of EC-EVs functional heterogeneity and specific regulatory mechanisms will provide novel insights and experimental basis for precision-targeted therapy of heart failure.

As crucial components of the cardiac microenvironment, IC-EVs exhibit complex dual regulatory roles in HF pathogenesis, working in concert with EVs from cardiomyocytes, cardiac fibroblasts, and endothelial cells to modulate HF progression. Experimental studies have demonstrated elevated levels of leukocyte-derived EVs in chronic HF patients compared to controls, with lymphocyte- and neutrophil-derived EVs being specifically detected in chronic heart failure (cHF). Ischemic cHF patients show significantly increased EVs production from T lymphocytes and neutrophils, with immune cell-derived EVs correlating with New York Heart Association (NYHA) functional class and being similarly activated in both heart failure with preserved ejection fraction (HFpEF) and heart failure with reduced ejection fraction (HFrEF) patients [5]. Gąsecka et al. [55] revealed that elevated leukocyte EVs concentrations are associated with a 4.7-fold increased risk of systolic dysfunction progression in heart failure with mid-range ejection fraction (HFmrEF) patients, indicating the pro-HF effects of IC-EVs across cHF subtypes. Additional studies demonstrated that M1 macrophage-derived exosomes can induce cardiomyocyte pyroptosis via miR-29a [56], while programmed death 1 (PD-1) inhibitor-treated macrophage exosomes promote cardiac senescence-related damage through the microRNA-34a-5p/PNUTS signaling pathway [57]. Conversely, M2 macrophage-derived EVs exhibit cardioprotective properties, with studies showing that exosomes carrying miR-148a and miR-378a-3p can reduce cardiomyocyte apoptosis and pyroptosis post-cardiac injury [58, 59]. These findings position IC-EVs as key mediators in HF’s complex pathological network and provide a theoretical foundation for developing immune-targeted precision therapies.

EVs are present in various biological fluids including blood, milk, and saliva, where they play important roles in long-distance cellular communication [60]. Although conventional wisdom suggests that fresh samples typically yield higher quantities of EVs, this is often difficult to achieve in practice. The isolation of EVs presents significant challenges due to their nanoscale size characteristics and frequent co-isolation with contaminants such as cellular debris and lipoproteins. Moreover, different isolation methods can substantially influence downstream analyses of EV cargo composition and physicochemical properties [61]. Current isolation techniques primarily include ultracentrifugation, ultrafiltration (also called microfiltration), precipitation, immunoaffinity capture, microfluidic technologies, and commercial kits [62, 63, 64, 65]. In terms of characterization, EVs analysis are advancing toward multidimensional precision profiling. EVs are typically characterized by their size, concentration, presence of protein markers, protein content, and other components. The main single-particle EVs analysis technologies currently available include nanoparticle tracking analysis, microscopy techniques (electron microscopy, cryo-electron microscopy, atomic force microscopy, and high-resolution microscopy), resistive pulse sensing, high-resolution flow cytometry, and Raman spectroscopy [66].

Current medical capabilities have yet to achieve large-scale clinical application of EVs, primarily due to technical challenges in EV isolation and production, as well as compositional variability [67, 68]. While production yields can be enhanced through various approaches including culture condition optimization (moderate pH adjustment, modulation of calcium ion concentration, glucose content, and oxygen levels), application of physical stimuli (temperature, light, or electrical pulses), or genetic modification to regulate EVs biogenesis-related proteins (e.g., Rab proteins and phospholipase D), the impacts of these interventions on EVs cargo composition, functionality, and stability require further investigation [69]. Realizing scalable applications necessitates comprehensive system-wide optimization: employing microcarriers with chemically defined media in cell culture to increase yield while preventing contamination; integrating size-exclusion chromatography with immunoaffinity capture during purification for efficient removal of lipoproteins and cellular debris; and establishing modular bioreactors with integrated “culture-purification-quality control” workflows to ensure EVs functional stability, thereby laying the foundation for large-scale production. This holistic approach not only addresses current technical limitations but also provides a clear pathway toward standardized, high-quality EVs manufacturing for clinical applications.

Currently, traditional biomarkers such as brain natriuretic peptide and N-terminal pro-brain natriuretic peptide are widely used in clinical diagnosis of HF, yet their limitations including insufficient sensitivity for early diagnosis and significant interference from non-cardiac factors remain unresolved [2, 70, 71]. Consequently, there is growing interest in identifying biomarkers with superior sensitivity and specificity to optimize HF clinical management. As key mediators of intercellular communication, EVs carry nucleic acid molecules (e.g., miRNAs, lncRNAs) which directly reflect the physiological or pathological status of their parent cells, emerging as a promising research focus for HF diagnosis, prognosis evaluation, and disease monitoring [72]. Wu et al. [73] demonstrated upregulated expression of EV-associated miR-92b-5p in serum from acute HF patients with dilated cardiomyopathy, suggesting its potential as a diagnostic biomarker for AHF secondary to this condition. Studies have confirmed significantly elevated levels of exosomal miR-192-5p and miR-320a in patients with acute decompensated HF compared to healthy controls. These miRNAs show positive correlations with patient age and cardiac chamber dimensions, exhibiting negative correlations with left ventricular ejection fraction and fractional shortening, which collectively indicate their capacity to reflect HF progression through associations with cardiac functional parameters [74, 75]. Notably, EV-encapsulated miR-126 and miR-199a demonstrate direct clinical relevance to cardiovascular outcomes, unlike their free circulating counterparts which show no clear association with HF-related events, highlighting the unique prognostic value of EV-packaged miRNAs for precise patient stratification and personalized intervention [76]. Furthermore, EVs characteristics including concentration, size distribution, and zeta potential, show diagnostic and prognostic potential in HF. Specifically, EVs concentration correlates strongly with overall survival in HF patients, where lower concentrations predict poorer outcomes, suggesting the utility of EVs profiling for prognostic assessment [77].

Beyond their utility as HF biomarkers, EVs possess inherent therapeutic advantages including excellent biocompatibility, low immunogenicity, and the ability to cross biological barriers, making them ideal nanocarriers for HF treatment. Through multifunctional design strategies and engineered modifications, EVs can be further optimized to enhance targeting efficiency, payload capacity, and therapeutic efficacy, thereby providing critical support for reversing myocardial remodeling and improving clinical outcomes in HF.

The conjugation of ischemic myocardium-targeting peptides or cardiac homing peptides to EVs surfaces enhances their specific targeting capability toward ischemic or injured myocardium, enabling efficient delivery of therapeutic molecules such as miRNAs to reduce apoptosis and restore cardiac function. Studies demonstrate that hypoxia-preconditioned BM-MSCs-derived EVs enriched with miR-125b-5p, when coupled with ischemic myocardium-targeting peptides and administered intravenously, these achieve highly specific delivery of cardioprotective miRNAs to the ischemic myocardium [78, 79, 80]. Similarly, cardiac homing peptide-modified stem cell-derived EVs facilitate targeted delivery of miR-21 gene therapy, improving cardiac functional recovery post myocardial infarction while attenuating subsequent heart failure development [81]. Mentkowski and Lang [82] experimentally verified that fusion of cardiomyocyte-specific binding peptides (CMPs) with the exosomal transmembrane protein LAMP2b significantly enhanced cardiomyocyte-specific uptake in vitro. Following intramyocardial injection in vivo, CMP-modified exosomes markedly reduced cardiomyocyte apoptosis, suggesting their therapeutic potential for heart failure treatment [52].

Functionalized EVs loaded with therapeutic molecules (e.g., antifibrotic drugs, pro-angiogenic factors, or nucleic acids) enable precise targeting of key pathways in HF progression. Ma et al. [83] demonstrated that exosomes derived from miR-132 mimic-treated BM-MSCs enhance angiogenesis and improve cardiac function in MI mice through targeted suppression of RASA1. Similarly, engineered EVs from cardiosphere-derived cells, which are naturally enriched with cardioprotective nucleic acids (miR-4488, miR-92a), attenuate atrial fibrosis-related conduction abnormalities and electrical dyssynchrony when administered intravenously in HFpEF rats. This effect is mediated through coordinated modulation of pro-inflammatory, pro-fibrotic, and anti-angiogenic pathways, resulting in significantly reduced atrial fibrillation incidence [84].

EVs derived from genetically modified or specific stem cell sources, which inherently carry therapeutic molecules with enhanced bioactivity, have been demonstrated to play significant roles in promoting angiogenesis and improving cardiac function. Studies have shown that adipose-derived mesenchymal stem cell exosomes (ADSC-Exos) can mitigate cardiac injury and enhance functional recovery in myocardial infarction (MI) mice through the miR-205 signaling pathway, suggesting the considerable therapeutic potential of miR-205-enriched ADSC-Exos for treating MI-induced heart failure [85, 86]. Comparative investigations revealed that exosomes isolated from Akt-modified human umbilical cord-derived mesenchymal stem cells induced more pronounced improvements in both angiogenesis and cardiac function than their unmodified counterparts, highlighting the superior therapeutic efficacy of genetically engineered EVs [87].

However, the clinical translation of engineered EVs remains constrained by multiple practical challenges, necessitating a more objective and rigorous perspective to evaluate their development prospects. In terms of core therapeutic efficacy, insufficient targeting specificity continues to be an unbroken bottleneck, while surface-anchored targeting peptides can enhance tropism toward damaged myocardium, complex circulatory dynamics and physiological microenvironments in vivo often lead to ligand degradation, shedding, or nonspecific binding to non-target organs. Some modified EVs still exhibit high off-target rates, which not only reduces drug concentration at lesion sites but may also increase potential risks in non-target tissues. Regarding industrial implementation, exorbitant production costs severely limit large-scale applications: The engineering processes involving ligand modification and precise therapeutic molecule loading require highly sophisticated equipment and technical precision. Moreover, the low efficiency and limited yield of high-purity engineered EVs result in per-dose treatment costs far exceeding conventional drugs, making them unsuitable for clinical batch production. On regulatory and safety fronts, the absence of clear approval standards and unknown long-term toxicity pose dual obstacles [88]. The clinical translation of engineered EVs therapies is hindered by the lack of standardization in classification, quality control, and regulatory approval. Additionally, critical long-term safety data remain scarce. Potential issues such as genomic integration risks associated with gene-edited EVs, immune memory reactions triggered by heterologous modified ligands, and toxic side effects from prolonged molecular accumulation in patients—most existing research focuses on short-term animal experiments rather than long-term clinical follow-up evidence. In addition, there are few clinical cases at present, and no large-scale, multi-center randomized controlled trial has been conducted to confirm its efficacy and safety in HF patients with different subtypes and stages.

Despite current challenges including scalable production, long-term safety validation, and limited clinical application, engineered EVs are poised to emerge as a novel therapeutic modality for reversing myocardial remodeling and improving HF prognosis as design technologies mature, thereby ushering in a new era of precision medicine for heart failure.