Extracellular vesicle-mediated communication in diabetic cardiomyopathy: cellular mechanisms, metabolic remodeling, and emerging therapeutic strategies

Literature search strategy and evidence classification

This narrative review is based on a comprehensive literature search of the PubMed, Web of Science, and Scopus databases through June 2026 using combinations of the keywords “diabetic cardiomyopathy”, “extracellular vesicles”, “small extracellular vesicles”, “exosomes”, “microRNAs”, and “long non-coding RNAs.” Although substantial mechanistic insights into EV-mediated cardiac remodeling have been obtained from diverse cardiovascular models, including HF, pressure overload, ischemia-reperfusion, and dilated cardiomyopathy models, such findings should not be considered equivalent to direct evidence in DCM. To facilitate interpretation of the available evidence, studies were classified according to an evidence hierarchy based on experimental context. Level 1 represents human clinical evidence derived from patient samples or cohorts; Level 2 includes in vivo DCM models with confirmed cardiac dysfunction; Level 3 includes non-diabetic cardiovascular disease models providing indirect mechanistic support that may be relevant to DCM but requires validation in diabetic settings; and Level 4 refers to in vitro studies, including high-glucose-treated cardiomyocytes or other relevant cellular models. Throughout this review, mechanistic conclusions regarding DCM are primarily based on evidence from diabetic models with confirmed cardiac phenotypes, whereas findings from non-DCM models are interpreted as supportive or hypothesis-generating evidence rather than established DCM-specific mechanisms. Because available evidence varies among EV cargoes and biological pathways, not all levels are represented in every table.

Overview of EVs

EVs have historically been classified according to their biogenesis pathways, physical characteristics, and size distribution. However, because current isolation methods and molecular markers cannot reliably distinguish EV subtypes based solely on biogenesis origin, contemporary EV research increasingly favors operational terminology. EVs may therefore be described using terms such as sEVs and medium/large EVs (m/lEVs) based on experimentally defined characteristics rather than as discrete biological entities. Although exosomes are commonly described as endosome-derived vesicles, definitive identification of bona fide exosomes remains challenging because they cannot be reliably distinguished from other sEV populations using currently available approaches.

EVs carry a diverse repertoire of bioactive molecules, including proteins, lipids, messenger RNAs (mRNAs), microRNAs (miRNAs), long non-coding RNAs (lncRNAs), and circular RNAs (circRNAs). This molecular cargo enables EVs to mediate intercellular communication, transferring molecular information from donor to recipient cells. Following interaction with recipient cells, EVs may fuse with the plasma membrane or undergo endocytosis, thereby releasing their cargo into the cytoplasm to regulate gene expression, intracellular signaling, and various cellular processes, including proliferation, differentiation, apoptosis, immune responses, angiogenesis, and tissue repair^[10].

The biological functions of EVs are influenced by multiple factors, including their biogenesis, cargo-sorting mechanisms, membrane composition, and uptake efficiency by recipient cells^[11]. In addition, extrinsic factors, including receptor availability and enzymatic activity in the extracellular milieu, further modulate EV function. Notably, a substantial proportion of circulating miRNAs is encapsulated within EVs, with some miRNA species also associating with high-density lipoproteins, underscoring the central role of EVs in systemic miRNA transport^[12]. Collectively, these characteristics establish EVs as important contributors to intercellular communication and physiological homeostasis. As comprehensively reviewed by Abbas et al., EVs derived from different cellular sources participate in the pathogenesis of diabetes and its complications, including cardiomyopathy, through the transfer of bioactive cargoes involved in inflammation, oxidative stress, and tissue remodeling^[13].

In DCM, EV-mediated intercellular communication orchestrates crosstalk among cardiomyocytes, endothelial cells, and fibroblasts [Figure 1]. Several miRNAs, although not yet experimentally validated as EV cargo in DCM, including miR-1, miR-223, miR-29 and miR-103, have been implicated in diabetes-associated cardiac remodeling and metabolic regulation. By shuttling these signaling molecules between cells, EVs can modulate recipient cell activity and contribute to the initiation and progression of DCM^[9]. Conversely, EVs derived from multiple sources, such as MSCs and macrophages, have shown therapeutic promise in cardiovascular disorders, including DCM^[11]. Accordingly, EVs are regarded as promising candidates for innovative diagnostic biomarkers and therapeutic approaches in DCM management^[14].

Figure 1. Evidence-based classification of EV-associated and EV-independent miRNA mechanisms in diabetic cardiomyopathy. Diabetic metabolic stress, including hyperglycemia, insulin resistance, lipotoxicity, and chronic inflammation, alters intercellular communication and miRNA-associated signaling in the diabetic heart. I, directly supported EV cargoes: miR-320 is transferred from diabetic cardiomyocytes to endothelial cells through EVs, where it suppresses IGF-1, HSP20, and ETS2 and impairs endothelial proliferation, migration, and tube formation. II, EV-associated signals with context-limited evidence: circulating miR-16-2-3p has been associated with diabetic coronary microvascular dysfunction, whereas dendritic-cell-derived miR-494-3p promotes angiogenesis in a myocardial infarction model; these findings should not be directly extrapolated to DCM. III, DCM-associated miRNA pathways without confirmed EV involvement: miR-34a, miR-200b/c, miR-195, and miR-451 contribute to vascular, fibrotic, hypertrophic, or metabolic abnormalities in diabetic or related cardiac models, but direct EV loading or EV-mediated transfer has not been established. The lower panel summarizes cellular processes potentially affected by EV-mediated communication and their contribution to diabetic cardiac remodeling and dysfunction. This evidence-based classification distinguishes experimentally supported EV cargoes from EV-associated signals and miRNA mechanisms for which direct EV involvement remains unconfirmed. ↑ Indicate upregulated; ↓ indicate downregulated; Solid arrows indicate established direct interactions, whereas dashed arrows denote putative or indirect regulatory relationships that require further experimental validation; T-bars represent inhibition. ACADM: Acyl-CoA dehydrogenase medium chain; AMPK: AMP-activated protein kinase; DCM: diabetic cardiomyopathy; DM: diabetes mellitus; EVs: extracellular vesicles; ETS2: ETS proto-oncogene 2; HG: hyperglycemia; HSP20: heat shock protein 20; IGF-1: insulin-like growth factor 1; LKB1: liver kinase B1; miRNA: microRNA; NO: nitric oxide; PK: pyruvate kinase; SIRT1: sirtuin 1; T2DM: type 2 diabetes mellitus; TGF-β: transforming growth factor beta.

Despite growing interest in the role of EVs in disease pathogenesis, fundamental aspects - including biogenesis, cargo loading, target cell specificity, and internalization - remain incompletely understood. Addressing these knowledge gaps will be essential for translating EV biology into clinically applicable diagnostic and therapeutic strategies. Although recent studies have substantially expanded our understanding of EV-mediated signaling, further investigation is required to define their precise contributions to DCM pathogenesis and their translational potential.

Pro-oxidant miRNAs and their pathogenic mechanisms

Chronic metabolic stress in obesity and type 2 diabetes alters the composition and functional properties of circulating EVs, which, upon uptake by cardiomyocytes, impair mitochondrial homeostasis and promote oxidative stress. For instance, adipocyte-derived sEVs from diabetic or high-fat-fed mice increase mitochondrial ROS production, depolarize the mitochondrial membrane, and promote cardiomyocyte injury - partly through transfer of miR-130b-3p, which suppresses peroxisome proliferator-activated receptor gamma coactivator 1-alpha (PGC-1α) expression^[31]. In hyperglycemic conditions, cardiomyocyte-derived EVs enriched with miR-320 are transferred to endothelial cells, where they suppress proliferation, migration, and tube formation, thereby impairing angiogenesis^[32]. Similarly, although miR-340-5p has not been specifically demonstrated to be EV-associated, it has been studied in diabetic heart tissue and cardiomyocytes. This miRNA directly targets the anti-apoptotic protein Mcl-1, causing mitochondrial dysfunction, ROS accumulation, and cardiomyocyte apoptosis in db/db mice. Inhibition of miR-340-5p reduces ROS levels and ameliorates fibrosis and cardiac dysfunction^[33].

Moreover, the miR-181 family exhibits context-dependent effects in the heart. miR-181a/b targets cytosolic proteins such as Phosphatase and Tensin Homolog (PTEN) (thereby influencing PI3K/AKT signaling), whereas miR-181c localizes to mitochondria and targets MT-CO1, resulting in distinct effects on mitochondrial function and infarct size^[34]. These findings highlight that the biological effects of miRNAs cannot be interpreted solely based on expression changes but require consideration of cellular localization, disease context, and experimental models.

Antioxidant miRNAs and miRNA-mediated protective pathways

In DCM, multiple miRNAs have been implicated in modulating oxidative stress and mitochondrial function. The miR-199a/214 cluster is notably upregulated under hypoxic or stress conditions in HF, where it directly suppresses peroxisome proliferator-activated receptor δ (PPARδ). This repression shifts cardiac metabolism away from mitochondrial fatty acid oxidation toward enhanced glycolysis. In vivo silencing of miR-199a/214 in stressed hearts restores mitochondrial β-oxidation of FFA and improves cardiac pump function and structure^[35]. miR-22 is downregulated in diabetic hearts; its overexpression reduces ROS production, decreases lipid peroxidation, as indicated by malondialdehyde (MDA) levels, restores superoxide dismutase (SOD) activity, and attenuates cardiomyocyte apoptosis through upregulation of sirtuin 1 (SIRT1), thereby mitigating high-glucose-induced oxidative injury^[36]. Similarly, miR-92a-2-5p is downregulated in diabetic hearts and high-glucose-stimulated cardiomyocytes; its therapeutic overexpression attenuates ROS accumulation, restores glutathione levels, lowers MDA, and reduces apoptosis by targeting mitogen-activated protein kinase (MAPK)-interacting kinase 2 (MKNK2) and suppressing p38-MAPK pathway phosphorylation^[37]. miR-144 has been implicated in oxidative stress regulation, but its role in DCM remains controversial. Reduced expression has been reported in diabetic cardiac models, whereas other studies suggest it modulates nuclear factor erythroid 2-related factor 2 (NRF2)-related antioxidant responses. Inhibition of miR-144 may hold therapeutic potential, yet whether its upregulation is pathogenic or compensatory is unclear^[38]. Similarly, miR-221 alleviates oxidative stress-induced remodeling via the p27/CDK2/mTOR axis, though this evidence derives primarily from HF models and requires validation in DCM^[39]. Another noteworthy regulator is miR-503, which participates in adaptive antioxidant responses. In a DCM rat model, miR-503 was significantly upregulated, while its direct target Nrf2 and downstream antioxidant enzymes were downregulated. Treatment with the Phase II enzyme inducer CPDT (5,6-dihydrocyclopenta-1,2-dithiole-3-thione) suppressed miR-503 and activated the NRF2/antioxidant response element (ARE) pathway, thereby mitigating oxidative stress and apoptosis in diabetic hearts^[40]. These findings suggest that increased miR-503 may represent a maladaptive response that suppresses endogenous antioxidant defense rather than a direct protective mechanism.

Collectively, modulation of oxidative stress-related miRNAs represents a potential strategy for attenuating DCM progression. For example, miR-141 has been reported to regulate mitochondrial phosphate carrier SLC25A3 in diabetic hearts, leading to impaired mitochondrial ATP production and contributing to diabetic cardiac dysfunction^[41], whereas tauroursodeoxycholic acid (TUDCA) treatment suppresses miR-340-5p to preserve mitochondrial homeostasis^[33], ultimately attenuating DCM progression.

ERS

In DCM, ERS acts as an important pathogenic pathway linking glucolipotoxicity, oxidative stress, inflammation, and cardiomyocyte apoptosis. Accumulating evidence indicates that diverse classes of non-coding RNA species (miRNAs, lncRNAs, and circRNAs) modulate DCM progression through regulation of the unfolded protein response (UPR) pathway. For example, lncRNA H19 is downregulated in diabetic hearts; restoring its expression attenuates ERS-induced C/EBP homologous protein (CHOP) upregulation and apoptosis, while concomitantly activating the PI3K/AKT/mTOR axis to promote cardiomyocyte survival^[42]. Levels of miR-30c are diminished in animal models of diabetes; overexpression of this miRNA suppresses PPARα activity via PGC-1β targeting, which in turn modulates ERS and autophagy to mitigate DCM^[43]. Members of the miR-30 family, along with miR-214, directly target X-box binding protein 1 (XBP1), a master transcriptional regulator of the UPR, thereby influencing cardiomyocyte adaptation and survival under metabolic stress^[44]. circRNAs have emerged as potential regulators of cardiovascular stress responses. For example, circDLGAP4 was reported to modulate endothelial dysfunction through the miR-143-HECTD1 axis in an ischemic stroke model, suggesting a possible role for circRNA-mediated regulation in cardiovascular remodeling. However, whether similar mechanisms contribute to ERS or DCM remains to be determined^[45]. Several miRNAs, including miR-133a^[46], have been reported to dampen ERS by downregulating the GRP78/CHOP/eIF2α cascade in various cardiovascular models. Despite these advances, direct evidence from well-characterized DCM models remains limited, and the precise mechanisms involved require further investigation.

Autophagy

Because prolonged ER stress is a well-recognized trigger of autophagy, these two processes constitute an interconnected stress-adaptation network that critically influences cardiomyocyte survival under diabetic conditions. Autophagy represents an essential adaptive response in DCM, yet its functional role is inherently dualistic. At moderate levels, autophagy maintains cellular homeostasis by clearing damaged organelles and protein aggregates, alleviating oxidative burden, and preserving myocardial energy metabolism. Conversely, when autophagic flux becomes excessive or dysregulated, enhanced degradation of essential cellular components may contribute to mitochondrial dysfunction and cardiomyocyte loss, thereby worsening contractile dysfunction and fibrotic remodeling^[47,48]. This delicate equilibrium is governed by key nutrient-sensing pathways, primarily mTOR and AMPK, whose outputs are further fine-tuned by a multilayered ncRNA regulatory network that integrates metabolic and stress cues to shape autophagic responses^[49].

A growing body of evidence has identified multiple non-coding RNAs (ncRNAs) as important modulators of autophagy in DCM. The lncRNA H19, for instance, has been shown to restrain excessive autophagy by attenuating DIRAS3-mediated inhibition of mTOR signaling, thereby preserving cardiac structure and function under hyperglycemic conditions^[49]. In contrast, lncRNA DCRF accelerates disease progression by functioning as a competitive endogenous RNA (ceRNA) for miR-551b-5p, leading to PCDH17 upregulation and heightened autophagic activity^[50]. Similarly, diminished lncRNA GAS5 expression activates the miR-221-3p/p27/mTOR axis, driving maladaptive autophagy and adverse cardiac remodeling^[51]. The pathogenic role of miR-221 in DCM has been further substantiated by studies demonstrating that angiotensin II exacerbates diabetic cardiac hypertrophy through activation of the JNK/c-Jun/miR-221 axis. This axis directly suppresses p27 and inhibits downstream autophagy^[52]. Conversely, pharmacological intervention with luteolin has been shown to attenuate DCM by suppressing JNK/c-Jun-regulated miR-221 and reversing autophagy inhibition^[53]. Emerging evidence suggests that additional ncRNAs, including circRNAs, may participate in the regulation of autophagy and stress responses, potentially extending autophagic signaling beyond individual cardiomyocytes to the broader diabetic cardiac microenvironment; however, these mechanisms remain incompletely characterized^[54].

Collectively, these observations underscore that autophagy in DCM acts as a highly dynamic, context-dependent process - protective at physiological levels but destructive when unrestrained. The expanding knowledge of ncRNA-mediated regulation offers a potential therapeutic avenue to precisely modulate autophagic flux. Although ncRNA-mediated regulation of autophagy represents a promising therapeutic strategy, defining the optimal timing, magnitude, and disease stage for intervention remains a major challenge. Validation in well-phenotyped preclinical models and prospective clinical studies is therefore essential^[47,48,54].

Cardiomyocyte apoptosis

Persistent ER stress and dysregulated autophagy ultimately converge on apoptotic signaling pathways, making cardiomyocyte apoptosis a major downstream consequence of chronic metabolic injury in DCM. This progressive loss of cardiomyocytes contributes directly to myocardial dysfunction and adverse ventricular remodeling. The metabolic derangements characteristic of diabetes, including impaired substrate utilization, lipotoxicity, ceramide accumulation, and mitochondrial dysfunction, promote excessive ROS generation and activation of caspase-dependent apoptotic pathways, ultimately tipping the balance toward cardiomyocyte death. Experimental studies have shown that apoptotic activity is markedly increased in diabetic myocardium compared with healthy hearts, thereby contributing to contractile dysfunction and adverse ventricular remodeling^[55].

Noncoding RNAs, including lncRNAs and miRNAs, form a complex regulatory network that influences apoptotic susceptibility in DCM [Table 1]. Many upregulated lncRNAs under hyperglycemic stress act as ceRNAs: for example, MALAT1 sponges miR-181a-5p, thereby modulating p53 and the EZH2/miR-22/ABCA1 axis^[72]; MIAT binds miR-22-3p, thereby upregulating DAPK2^[73]; KCNQ1OT1 competes for miR-181a-5p, leading to elevated programmed cell death 4 (PDCD4) levels^[74]; and nuclear paraspeckle assembly transcript 1 (NEAT1) acts as a decoy for miR-140-5p, resulting in increased histone deacetylase 4 (HDAC4) and reduced NRF2^[75]. In contrast, downregulation of lncRNA H19 in DCM diminishes its processing into miR-675, thereby lifting repression of voltage-dependent anion channel 1 (VDAC1) and promoting mitochondrial permeability transition, inflammatory signaling, and apoptosis^[76]. Increased miR-34a expression has been associated with reduced SIRT1 signaling and enhanced apoptotic and fibrotic responses in diabetic cardiac models^[60,77]. Accordingly, inhibition of miR-34a has shown protective effects in experimental settings^[60,77]. However, the biological interpretation of miR-34a requires caution, as its functional consequences may vary according to disease stage, cellular origin, and experimental intervention strategy. Notably, the long noncoding RNA HOX transcript antisense RNA (HOTAIR) can function as a molecular sponge for miR-34a, thereby relieving miR-34a-mediated translational repression of SIRT1 and attenuating oxidative stress-induced cell death^[60]. Similarly, miR-22 upregulates SIRT1 to counteract injury^[36], and miR-29a targets Bak1 to blunt apoptotic signaling in high-glucose or Streptozotocin (STZ)-induced models^[78].

Beyond intracellular ncRNA regulation, EV-mediated intercellular transfer of ncRNAs provides an additional mechanism by which apoptosis is propagated or restrained within the diabetic myocardium. For instance, miR-21, which is enriched in circulating exosomes, has been identified as a novel biomarker for the diagnosis of subclinical DCM^[79]. Furthermore, cardiac progenitor cell-derived exosomal miR-21 has been shown to protect cardiomyocytes from apoptosis by targeting PDCD4, as shown in models of oxidative stress and cardiac injury^[56]. These findings support a potential cardioprotective role of miR-21-containing EVs in specific experimental contexts; however, miR-21 may exert distinct effects depending on EV origin and disease setting. In contrast, cardiomyocyte-derived EVs with elevated levels of miR-320 contribute to functional deterioration by impairing endothelial repair and angiogenic capacity^[32]. Consistent with this observation, inhibition of miR-320 reduces apoptosis and improves cardiac function in experimental diabetes, suggesting that miR-320 contributes to DCM progression^[80]. miR-25 has also been shown to protect cardiomyocytes against oxidative damage by directly targeting the mitochondrial calcium uniporter (MCU), thereby reducing mitochondrial Ca²⁺ overload and inactivating the mitochondrial apoptosis pathway^[81].

In addition to ncRNAs, several additional miRNAs have emerged as significant modulators of cardiomyocyte fate in DCM, often operating at the interface of apoptosis and other stress responses. miR-483-3p is upregulated in STZ-induced diabetic mice and high-glucose-treated cardiomyocytes, where it promotes apoptosis under glucose stress^[66]. Members of the miR-30 family, notably miR-30c, modulate the p53-p21 axis to influence hypertrophic and apoptotic responses in high-glucose conditions^[82], whereas miR-30d has been shown to regulate pyroptosis through direct targeting of FOXO3a, thereby linking inflammasome-driven cell death to diabetic cardiac injury^[83]. Therapeutic inhibition of miR-195 in STZ mice attenuates cardiac hypertrophy and reduces apoptosis, revealing a deleterious role for miR-195 in DCM^[61]. The cardiac-enriched miR-378 modulates the IGF-1/IGF receptor (IGF-1R) signaling axis, linking metabolic remodeling with cardiomyocyte survival^[84]. Finally, miR-1 controls apoptosis in part through modulation of liver X receptor α (LXRα) and mitochondrial ROS pathways in glucose-stressed cardiomyocytes, illustrating how miRNAs can integrate nuclear receptor signaling with mitochondria-dependent cell death^[85]. Collectively, these findings highlight that both intracellular and EV-associated ncRNAs converge on key apoptotic signaling pathways - including p53, SIRT1, IGF-1/IGF-1R, BAK1, FOXO3, and CD36/PPARγ - to regulate cardiomyocyte survival in DCM. Given the context-dependent functions of several miRNAs, future therapeutic strategies will require careful consideration of RNA species, cellular origin, disease stage, and delivery approach. These shared molecular nodes may represent promising therapeutic targets for limiting myocardial injury and delaying DCM progression.

EV-mediated microvascular dysfunction and angiogenic regulation in DCM

Coronary microvascular dysfunction is a core pathological feature of DCM, particularly in the absence of obstructive epicardial coronary artery disease. Clinical imaging investigations have demonstrated that reduced coronary flow reserve (CFR) is significantly associated with adverse outcomes in patients with diabetes, underscoring the prognostic significance of coronary microvascular impairment^[134]. Hyperglycemia initially damages CMECs, provoking nitric oxide-dependent vasoregulatory defects and barrier disruption. Notably, under acute nutrient stress, cardiomyocytes increase EV secretion and selectively load pro-angiogenic cargo capable of transiently stimulating endothelial migration and tube formation; however, with sustained hyperglycemia, the EV cargo profile shifts toward fibrotic and anti-angiogenic signals that compromise microvascular integrity and perfusion^[135]. CMEC-derived EVs are critical intermediaries in this process: under high-glucose conditions, they become enriched in TGF-β1 mRNA, which is delivered to cardiomyocytes and fibroblasts, promoting perivascular fibrosis and abnormal angiogenic responses^[120]. Because coronary microvascular dysfunction is increasingly recognized as a major contributor to myocardial ischemia, interstitial fibrosis, and the development of HFpEF in diabetes, preserving microvascular integrity has emerged as an important therapeutic objective in DCM.

The biological effects of EVs on the coronary microcirculation are therefore highly context-dependent and largely determined by their cellular origin and molecular cargo. In type 2 diabetic models, cardiomyocyte-derived EVs enriched in miR-320 have been shown to transfer this anti-angiogenic signal to endothelial cells, suppressing IGF-1, HSP20, and ETS proto-oncogene 2 (ETS2) expression and impairing endothelial proliferation, migration, and tube formation^[32]. Conversely, EV-based therapeutic strategies show context-dependent protective effects in experimental cardiovascular models. In experimental diabetic heart disease models, therapeutic inhibition of miR-320 improves cardiac function and attenuates adverse remodeling, although this effect reflects direct miR-320 suppression rather than anti-miR-320-loaded EV delivery^[80]. MSC-derived exosomal miR-21-5p has been shown to enhance cardiac contractility in human cardiac tissue models, but its therapeutic relevance to DCM requires further validation^[57]. EV-associated regulatory mechanisms have also been implicated in diabetic vascular dysfunction and myocardial repair. Circulating exosomal miR-16-2-3p has been associated with diabetic coronary microvascular dysfunction through regulation of endothelial fatty acid metabolism^[58], whereas dendritic-cell-derived exosomal miR-494-3p promotes angiogenesis after myocardial infarction rather than DCM^[59]. HSP20 has been reported to interact with EV biogenesis machinery and enhance cardiomyocyte-derived EV secretion under diabetic conditions, with associated cardioprotective effects; however, its identity as an EV cargo remains unconfirmed^[112].

Among endothelial regulatory miRNAs, the miR-200 family contributes to microvascular pathology by driving EndMT and impairing vasodilatory function: miR-200b promotes EndMT via p300-dependent transcriptional reprogramming and collagen deposition in diabetic hearts, whereas miR-200c is upregulated in diabetic arteries and suppresses zinc finger E-box binding homeobox 1 (ZEB1), resulting in enhanced cyclooxygenase-2 (COX-2)/prostaglandin E2 (PGE2) signaling and diminished endothelium-dependent vasodilation - mechanisms that collectively accelerate microvascular deterioration^[64,65].

Recent studies have identified miR-550a-5p as an additional contributor to diabetic microvascular pathology. This miRNA, which is elevated in circulating EVs from patients with microvascular dysfunction, targets the Hippo-YAP/endothelial nitric oxide synthase (eNOS) axis, potentially mediating adverse endothelial-cardiomyocyte communication and contributing to myocardial injury^[136].

Although EV-targeted therapies have not yet entered clinical practice, findings from EV-independent interventions that improve coronary microvascular function provide important proof-of-concept for restoring vascular homeostasis in DCM. For example, modulation of non-EV pathways that influence microvascular tone also carries translational relevance: mineralocorticoid receptor antagonism with spironolactone improved CFR in individuals with type 2 diabetes, providing support for therapeutic strategies that target microvascular dysfunction in diabetes^[137].

Overall, current evidence indicates that EVs exert divergent effects in DCM depending on their cellular origin and molecular cargo. Pathological EVs released from metabolically stressed cardiac cells may contribute to microvascular injury and adverse remodeling, whereas therapeutic EV platforms have shown potential to deliver protective signals [Table 3].

Beyond microvascular regulation, the bidirectional effects of EV-mediated communication extend across multiple pathological domains of DCM, including metabolic dysfunction, oxidative stress, apoptosis, hypertrophic remodeling, and fibrosis. As summarized in Figure 2, pathological EVs released from metabolically stressed cardiac cells transfer detrimental cargoes, such as miR-34a, miR-320, and MST1, thereby amplifying these maladaptive processes. Conversely, therapeutic EV-based approaches derived from MSCs or engineered platforms may exert protective effects by modulating metabolic homeostasis, vascular repair, and cardiac remodeling. However, the contribution of individual molecules requires careful interpretation because direct EV-mediated delivery has not been established for all candidate factors. This integrated framework highlights EV-mediated signaling as an important mediator of diabetic cardiac remodeling and provides a conceptual basis for understanding the diverse mechanisms underlying DCM progression.

Figure 2. EV-mediated intercellular communication in DCM: pathological mechanisms and therapeutic opportunities. The diabetic milieu characterized by hyperglycemia and IR (center) alters EV release and cargo composition from cardiac cells. EVs function as mediators of communication among cardiomyocytes, endothelial cells, and fibroblasts, thereby regulating multiple pathological processes involved in DCM progression. The left (red) pathway illustrates the pathogenic effects of EVs derived from metabolically stressed cells. These EVs transfer pathological molecular signals, including validated EV-associated cargoes and EV-related regulatory molecules, including miR-320 (cardiomyocyte-derived), miR-130b-3p (adipocyte-derived) and MST1 (CMEC-derived), which promote oxidative stress, fibrosis, apoptosis, microvascular dysfunction, and metabolic impairment, ultimately contributing to cardiac remodeling and HF. The right (green) pathway highlights protective/therapeutic EVs derived from MSCs or engineered platforms. These approaches deliver or modulate cardioprotective molecular signals, including validated EV-associated cargoes and protective regulators, such as miR-21-5p (CPC-derived exosomes), which enhance metabolic homeostasis, preserve vascular function, attenuate fibrosis and apoptosis, and improve cardiac remodeling. The balance between pathogenic and protective EV signaling influences the progression or attenuation of diabetic cardiac injury. Red ↑ indicate upregulated; green ↓ indicate downregulated; Solid arrows indicate established direct interactions, whereas dashed arrows denote putative or indirect regulatory relationships that require further experimental validation. CPC: Cardiac progenitor cell; DCM: diabetic cardiomyopathy; EndMT: endothelial-to-mesenchymal transition; ER: endoplasmic reticulum; EVs: extracellular vesicles; HF: heart failure; IR: insulin resistance; miRNA: microRNA; MSCs: mesenchymal stem cells; MST1: macrophage stimulating 1; CMEC: cardiac microvascular endothelial cell.

Emerging evidence for EV-mediated metabolic organ-heart communication

Beyond cardiac resident cells, EVs may represent an additional layer of communication between the diabetic heart and metabolically active peripheral organs. This concept is particularly relevant to DCM, which develops within the context of systemic metabolic disturbances rather than isolated myocardial abnormalities. Diabetes-induced remodeling of adipose tissue, liver, and skeletal muscle is accompanied by alterations in circulating EV abundance and cargo composition, raising the possibility that EVs contribute to metabolic organ-heart communication^[138-140]. However, compared with cardiac intercellular EV signaling, direct evidence establishing organ-specific EV transfer and functional effects in DCM remains limited.

The adipose tissue-heart axis represents one of the most extensively investigated examples of EV-mediated metabolic communication. Adipose-derived EV-associated miRNAs have been implicated in regulating glucose homeostasis, IR, and diabetes-related cardiovascular complications^[138]. Specifically, sEVs released from adipocytes of diabetic or high-fat-fed mice carry miR-130b-3p, which can be transferred to cardiomyocytes and suppress PGC-1α expression, thereby promoting mitochondrial oxidative stress and cardiomyocyte injury. These findings provide mechanistic evidence that metabolic abnormalities originating from peripheral tissues may influence myocardial homeostasis through EV-mediated signaling. Beyond intracellular cargo transfer, adipocyte-derived sEVs also convey protective signals through surface-associated adiponectin (APN), which interacts with cardiomyocyte AdipoR1 to promote metabolic homeostasis. In diabetes, G protein-coupled receptor kinase 2 (GRK2)-mediated phosphorylation of AdipoR1 disrupts this protective signaling pathway, potentially shifting the adipose-heart axis from an adaptive to a maladaptive communication network^[141]. Furthermore, recent evidence suggests that adipocyte-derived EVs can transfer oxidatively damaged mitochondrial components to cardiomyocytes, impairing myocardial bioenergetics and cell survival^[142]. Together, these findings indicate that diabetes-associated EV dysfunction may involve both altered cargo transfer and impaired surface-mediated signaling.

Other metabolically active organs may also participate in EV-mediated communication with the heart. Metabolic stress associated with hepatic steatosis alters EV release and molecular composition, and liver-derived sEVs have been implicated in inter-organ metabolic and inflammatory regulation^[143]. However, direct evidence linking hepatic EV-mediated communication to DCM remains limited. Although these observations suggest a potential liver-heart communication pathway, direct evidence linking hepatic EVs to DCM progression remains insufficient. Similarly, exercise-induced remodeling of skeletal muscle-derived EV profiles, including enrichment of cargoes related to mitochondrial function and angiogenesis, suggests a possible mechanism through which peripheral metabolic adaptation may influence cardiovascular health^[139]. However, whether skeletal muscle-derived EVs directly regulate myocardial function in diabetic hearts remains to be determined.

Collectively, EV-mediated metabolic organ-heart communication represents an emerging extension of the DCM framework, linking systemic metabolic dysregulation with myocardial injury. Nevertheless, current evidence for EV-mediated inter-organ communication remains substantially less established than that for cardiac cell-to-cell EV signaling. Future studies incorporating organ-specific EV tracing, functional validation, and longitudinal human cohorts will be essential to determine whether circulating EVs actively contribute to DCM pathogenesis or primarily serve as biomarkers reflecting systemic metabolic disturbances.

Conclusion

DCM represents a complex metabolic and inflammatory cardiac disorder driven by interconnected processes, including impaired substrate utilization, chronic inflammation, oxidative stress, microvascular dysfunction, and maladaptive structural remodeling. Over the past decade, EVs have emerged as important regulators of cellular communication, linking cardiomyocytes, endothelial cells, fibroblasts, macrophages, and metabolically active organs. Through their diverse molecular cargo, including miRNAs, lncRNAs, circRNAs, and proteins, EVs regulate pathways involved in insulin sensitivity, mitochondrial homeostasis, fibrosis, hypertrophic remodeling, cell survival, and endothelial integrity.

EV-associated molecular signatures may reflect dynamic alterations in disease states and therapeutic responses that are not fully captured by conventional biomarkers, supporting their potential utility for diagnosis, prognosis, and disease monitoring. Notably, a recent comprehensive review has systematically consolidated the theranostic potential of EV-mediated miRNAs in type 2 diabetes and its complications, highlighting their dual role as both non-invasive biomarkers for early disease detection and promising therapeutic targets^[146]. Preclinical proof-of-concept studies further indicate that MSC-derived and engineered EVs can ameliorate diabetic cardiac injury, providing a foundation for the future development of EV-based therapeutic approaches.

Despite these advances, the field remains in its early stages. Most available evidence remains associative rather than causal, and definitive links between EV cargo and human DCM pathogenesis remain to be established. Technical heterogeneity in EV isolation methods, nomenclature, and characterization criteria continues to limit reproducibility and cross-study comparisons. Whether EV alterations represent primary pathogenic mechanisms, secondary compensatory responses, or nonspecific consequences of metabolic stress remains unresolved. Furthermore, clinical translation of EV-based therapies faces substantial challenges, including manufacturing standardization, tissue-specific delivery, immunocompatibility, and long-term safety evaluation.

Advancing the field will require a transition from descriptive observations toward mechanistically validated and clinically relevant investigations. Key priorities include genetic or pharmacological modulation of EV biogenesis and secretion in DCM models; functional validation of EV cargoes in well-characterized experimental systems; establishment of large, deeply phenotyped clinical cohorts to define biomarker specificity; implementation of MISEV2023-aligned standards; and systematic investigation of how commonly used antidiabetic therapies remodel EV biology. For EV-based interventions to reach clinical application, they must meet rigorous standards for manufacturing consistency, biological activity, safety, and clinical efficacy comparable to other biologic therapies.

The therapeutic potential of EVs in DCM remains highly promising, yet substantial knowledge gaps persist. Bridging these gaps will require sustained research efforts, methodological rigor, and a comprehensive understanding of the complex molecular communication networks connecting EVs, individual cell populations, and metabolically active organs. Further clarification of EV-mediated communication may facilitate the development of improved biomarkers and therapeutic strategies for DCM.