Mitochondrial extracellular vesicles in cardiorenal syndrome: emerging mediators of mitochondrial signaling and inter-organ crosstalk

Abstract

Cardiorenal syndrome (CRS) comprises five clinically distinct patterns of acute, chronic, or systemic heart–kidney interaction. Mitochondrial dysfunction is shared by cardiac and renal injury, but the cardiorenal setting is distinguished by the possibility that mitochondrial stress arising in one organ is externalized and transmitted to the other. Extracellular vesicles (EVs) are established mediators of intercellular communication, and EVs carrying mitochondrial DNA, proteins, lipids, RNA, or structurally preserved mitochondrial material—collectively referred to here as mitochondrial extracellular vesicles (mitoEVs)—may connect mitochondrial quality control with systemic signaling. Direct CRS-specific evidence, however, remains limited: patient-derived studies support pathogenic effects of total circulating EVs, whereas most mitoEV-specific mechanisms are inferred from related cardiovascular, renal, inflammatory, cancer, or regenerative models. Accordingly, this review presents an evidence-graded conceptual framework rather than a definitive mechanistic summary. We define and classify mitoEVs, outline methodological requirements for their isolation, same-particle identification, cargo-topology analysis, quantification, and functional validation, and map the available evidence across the five CRS subtypes. We further propose the mitoEV–mitophagy–inflammation axis as a working hypothesis in which impaired mitochondrial quality control may favor vesicular export, inflammatory activation in recipient cells, and secondary mitochondrial dysfunction. Finally, we evaluate the biomarker and therapeutic potential of mitoEVs while emphasizing the need for subtype-specific clinical validation, standardized analytical workflows, source-resolved studies, and rigorous distinction between pathological and reparative vesicle populations.

1. Introduction

Cardiorenal syndrome (CRS) encompasses a spectrum of disorders characterized by bidirectional interactions between cardiac and renal dysfunction. According to the Acute Dialysis Quality Initiative classification, CRS includes type 1 (acute cardiac dysfunction leading to acute kidney injury), type 2 (chronic cardiac dysfunction leading to progressive renal impairment), type 3 (acute kidney injury leading to acute cardiac dysfunction), type 4 (chronic kidney disease leading to chronic cardiac injury or heart failure), and type 5 (a systemic disorder causing simultaneous cardiac and renal dysfunction). These subtypes differ in the initiating organ, temporal course, hemodynamic environment, inflammatory burden, and likely sources and kinetics of circulating signals. Hemodynamic disturbance, neurohormonal activation, venous congestion, inflammation, and oxidative stress remain central to CRS, but do not fully explain how cellular injury is propagated between distant organs [1,2].

Mitochondrial dysfunction is a shared pathological feature of cardiac and renal injury. Cardiomyocytes and renal tubular epithelial cells are highly energy-dependent and vulnerable to disturbances in oxidative phosphorylation, mitochondrial dynamics, redox homeostasis, and mitophagy [3,4]. The cardiorenal context is distinct from isolated cardiac or renal disease because mitochondrial injury occurs within a temporally and directionally coupled two-organ system: cardiac dysfunction may expose the kidney to hypoperfusion, congestion, neurohormonal activation, and circulating inflammatory mediators, whereas renal dysfunction creates a uremic, oxidative, and inflammatory environment that can impair cardiac metabolism and function. In type 5 CRS, both organs may be injured simultaneously and circulating vesicles may arise from multiple tissues and blood-cell populations. The central question is therefore not simply whether mitochondrial dysfunction occurs, but how mitochondrial stress may be converted into mobile signals that propagate or sustain heart–kidney injury.

Extracellular vesicles (EVs) are increasingly recognized as mediators of intercellular and inter-organ communication in cardiovascular and kidney disease. Circulating EVs isolated from patients with CRS induce renal epithelial and endothelial injury in a kidney-on-chip system, providing direct functional evidence for pathogenic circulating total EVs, although neither their organ of origin nor a mitoEV-specific contribution was resolved [5]. In CKD, EVs are also being investigated as biomarkers and mediators of inflammation, fibrosis, apoptosis, and oxidative stress [6]. Recent articles in Renal Failure have highlighted mitochondria-related kidney research and EV biology in diabetic kidney disease [4,7].

Beyond conventional EV cargo, experimental studies show that EVs can transport mitochondrial DNA (mtDNA), mitochondrial proteins, regulatory RNAs, lipids, and, in selected settings, structurally preserved mitochondrial material, thereby altering mitochondrial homeostasis in recipient cells [8]. These observations have stimulated interest in mitochondrial extracellular vesicles (mitoEVs), a heterogeneous operational category of EVs carrying mitochondrial constituents [9].

MitoEVs must be distinguished from intracellular mitochondria-derived vesicles (MDVs). MDVs bud selectively from mitochondrial membranes and primarily mediate intracellular quality control, whereas mitoEVs are extracellular particles detected after secretion [10]. The two pathways may intersect when mitochondrial cargo is redirected from intracellular degradation toward extracellular release, but they are not synonymous.

The functional consequences of mitochondrial cargo-containing EVs are context-dependent. Vesicles enriched in mitochondrial damage-associated molecular patterns may amplify inflammation, whereas stem cell-derived or engineered EVs can stabilize mitochondrial homeostasis and restore bioenergetics in experimental injury models [11,12]. Nevertheless, most CRS studies have analyzed total EV populations without resolving mitochondrial subtypes, and no mitoEV-specific causal pathway across the heart–kidney axis has been established.

In this review, we define and classify mitoEVs, summarize practical requirements for their identification and functional validation, examine mitochondrial dysfunction as a source of vesicular signaling, map the evidence across the five CRS subtypes, and discuss biomarker and therapeutic implications. We also propose a mitoEV–mitophagy–inflammation axis as a testable working hypothesis rather than an established CRS mechanism.

To distinguish evidence from biological plausibility, we use three evidence tiers. Tier 1 denotes direct evidence from patients or experimental models of CRS, with total EV evidence separated from mitoEV-specific evidence. Tier 2 denotes mechanistic evidence from related cardiovascular, renal, or systemic disease models. Tier 3 denotes theoretical extrapolation to CRS that has not been directly tested.

2. Definition and classification of mitoEVs

Mitochondrial extracellular vesicles (mitoEVs) are generally understood as extracellular vesicles that carry mitochondrial constituents, including mitochondrial proteins, lipids, mtDNA, mtRNA, or, in some settings, structurally preserved mitochondria. This definition is conceptually useful, but it should not be interpreted as implying a uniform vesicle population. Rather, mitoEVs represent a heterogeneous group of extracellular particles generated under diverse physiological and pathological conditions, ranging from conventional EV secretion to stress-induced organelle extrusion [8,9].

A careful definition of mitoEVs requires two distinctions. First, mitoEVs should be distinguished from total EVs. The current EV field recommends that vesicles be characterized by a combination of physical properties, molecular markers, and isolation procedures rather than by a single marker or size range alone [13]. Accordingly, a mitoEV preparation should ideally demonstrate both EV identity, such as enrichment of CD9, CD63, CD81, TSG101, or ALIX, and mitochondrial identity, such as TOMM20, TIM23, COX IV, VDAC1, ATP5A1, TFAM, cardiolipin, or mtDNA [8,13]. Second, mitoEVs should be distinguished from intracellular mitochondria-derived vesicles (MDVs). MDVs are generated by selective budding from mitochondrial membranes and function mainly as intracellular quality-control intermediates, whereas mitoEVs are extracellular particles that deliver mitochondrial cargo to the extracellular space or recipient cells [10]. These two processes may intersect, but they should not be treated as synonymous.

Given this heterogeneity, a classification based primarily on biogenesis is more defensible than a classification based on size alone. At least four biogenetic routes are currently relevant to mitoEV biology. The first is the endosomal or multivesicular body pathway, in which mitochondrial material is incorporated into EVs released after fusion of multivesicular bodies with the plasma membrane. A recent mechanistic study showed that lysosomal inhibition or Rab7 deficiency can promote secretion of mitochondria-containing large EVs. These vesicles carried EV markers such as ALIX, TSG101, CD81, and CD63 together with mitochondrial proteins including TOMM20, TIMM23, MT-CO1, and SOD2, and the authors concluded that mitochondrial secretion in EVs can serve as an alternative clearance route when lysosomal degradation is compromised [8,14]. This provides strong evidence that endosomal trafficking can generate mitochondria-containing EVs, particularly when canonical degradative pathways are impaired.

The second route is plasma membrane budding, which gives rise to microvesicle-type mitoEVs. This mechanism is especially relevant under cellular stress, when mitochondrial fragments, mtDNA, oxidized mitochondrial material, or even intact organelles may be redirected toward outward budding and extracellular release. In the context of systemic inflammation, this pathway is biologically important because microvesicle release can convert intracellular mitochondrial injury into extracellular inflammatory signaling. For a CRS-oriented review, this route is particularly relevant because sepsis, ischemia, uremic stress, and heart failure are all conditions in which mitochondrial damage and inflammatory vesicle release may coexist [15,16].

The third route is autophagy- or mitophagy-related secretion. When the degradative arm of mitophagy is incomplete or overwhelmed, mitochondrial cargo may be diverted toward noncanonical secretory routes. This concept remains less mature than the endosomal and microvesicle pathways, but it is important because it links mitoEV biology with mitophagy flux rather than simply with vesicle release. Recent studies describing secretory mitophagy and EV-mediated export of mitochondrial material support the idea that mitochondrial disposal can be shifted from intracellular degradation to extracellular release under selected stress conditions [17]. In CRS, where both cardiac and renal tissues exhibit mitochondrial stress and impaired quality control, this pathway provides a plausible mechanism through which failed mitophagy may become a source of circulating mitochondrial danger signals.

The fourth route comprises larger or more specialized mitochondrial extrusion mechanisms. Exophers were initially described as large extracellular vesicular structures that remove protein aggregates and damaged organelles from neurons, including oxidized mitochondria [18]. Migrasome-mediated mitocytosis has subsequently been reported as a mitochondrial quality-control process in migrating cells, in which damaged mitochondria are transported into migrasomes and disposed of outside the cell [19]. More recently, mitopherogenesis was described as a mitochondria-specific ectocytosis process during sperm development, generating vesicles termed mitophers that contain individual mitochondria [20]. These pathways are not yet established as major contributors to human CRS, but they expand the conceptual boundary of mitoEV biology by showing that mitochondrial export can occur through developmentally or context-specific vesicular mechanisms (Figure 1).

Figure 1.

Biogenesis and classification of mitoEVs.

Marker-based classification is useful for experimental reporting but should remain operational rather than taxonomic. A practical approach is to describe vesicle preparations as mitochondrial marker-positive EVs, for example TOMM20⁺CD63⁺, COX IV⁺CD81⁺, mtDNA⁺EVs, or ATP5A1⁺EVs, while specifying the isolation method and biological source. This avoids the misleading assumption that all mitoEVs share a single marker profile. Similarly, size-based nomenclature should be used cautiously. Small EVs and exosome-derived mitoEVs may overlap with MDVs in size, whereas microvesicle-type or organelle-containing mitoEVs may extend into the submicron or micron range. Therefore, size can support characterization, but it cannot define mitoEV subtype on its own [9,13].

For this review, ‘mitoEVs’ is used as an umbrella operational term only when EV identity and mitochondrial cargo have both been demonstrated. We classify such preparations primarily by proposed biogenesis—endosome/exosome-derived, plasma-membrane-derived, autophagy- or mitophagy-related, and specialized extrusion-associated mitoEVs—while avoiding taxonomic claims based on size or a single marker [9,14,19,20]. ‘Mitochondria-containing EVs’ is reserved for vesicles with structural evidence of intact or partial mitochondria; ‘mitochondrial cargo-containing EVs’ refers to EVs carrying mtDNA, mitochondrial proteins, lipids, or RNA without evidence of an intact organelle; and ‘mitochondria-enriched small EVs’ or ‘mt-sEVs’ is retained only when reproducing the terminology of an original study. These terms are not interchangeable.

MitoEVs are generated through multiple biogenetic pathways [1]. Exosome-derived mitoEVs originate from the endosomal system, where mitochondrial components are incorporated into early endosomes, subsequently forming multivesicular bodies (MVBs) and released as exosomes upon fusion with the plasma membrane [2]. Microvesicle-derived mitoEVs are formed by direct outward budding of the plasma membrane under cellular stress conditions [3]. Autophagy-related mitoEVs arise from secretory autophagy, in which autophagosomes or amphisomes fuze with the plasma membrane to release mitochondrial cargo [4]. Specialized extrusion pathways include exophers, migrasomes, and mitophers, which mediate the removal of damaged mitochondria or mitochondrial components through distinct vesicular structures. Collectively, mitoEVs represent a heterogeneous population carrying diverse mitochondrial cargo, including mitochondrial DNA (mtDNA), mitochondrial RNA (mtRNA), mitochondrial proteins (e.g. TOMM20, COX IV), cardiolipin, and metabolites. Abbreviations: MVB, multivesicular body; mtDNA, mitochondrial DNA; mtRNA, mitochondrial RNA; TOMM20, translocase of outer mitochondrial membrane 20; COX IV, cytochrome c oxidase subunit IV.

3. Methodological framework for mitoEV identification and functional validation

MISEV2023 provides the minimum general framework for EV nomenclature, separation, characterization, contaminant assessment, dose reporting, and functional studies, but it does not establish mitoEV-specific diagnostic criteria [13]. Operational identification therefore requires convergent evidence: EV identity, mitochondrial cargo, confirmation that both features occur within the same particle where feasible, and exclusion of non-vesicular mitochondrial material. For blood studies, collection, anticoagulant, processing delay, centrifugation, platelet depletion, hemolysis, storage, and freeze–thaw history should be reported in accordance with MIBlood-EV principles [21]. Urine studies should document collection conditions, volume or flow, creatinine, osmolality, proteinuria, storage, and uromodulin interference [22].

Because no single procedure selectively isolates all mitoEV subtypes, a staged enrichment and validation strategy is preferable. After standardized biofluid processing, EV-enriched fractions may be obtained by size-exclusion chromatography, density-gradient centrifugation, or an orthogonal combination of separation methods. Plasma preparations should be assessed for albumin, apolipoproteins, platelet-derived particles, and other non-EV contaminants; urine preparations should be assessed for protein aggregates, uromodulin, and non-vesicular nucleic acids. Immunoaffinity capture using TOMM20, VDAC1, an EV surface marker, or another accessible antigen may enrich a subpopulation but cannot independently define mitoEVs, because epitope accessibility depends on membrane topology and free mitochondrial fragments may also be captured.

EV identity should be supported by complementary physical and molecular approaches, and mitochondrial identity by more than one mitochondrial constituent, such as TOMM20, TIMM23, COX IV, VDAC1, ATP5A1, TFAM, cardiolipin, mtDNA, or mtRNA. Bulk co-detection of an EV-associated marker and a mitochondrial marker does not demonstrate that both are present on the same particle. Nano-flow cytometry, imaging flow cytometry, single-particle immunoassays, super-resolution imaging, or immunogold electron microscopy should therefore be used for same-particle confirmation where technically feasible. Protease- or nuclease-protection assays, performed with and without membrane permeabilization, can distinguish externally associated material from protected intravesicular cargo.

Quantification should report more than one denominator, including total particle number, EV-marker-positive particles, mitochondrial-marker-positive EVs, and mitochondrial cargo per particle or per EV-marker-positive particle. Urinary measurements should additionally be normalized using complementary measures such as urinary creatinine, osmolality, and, where available, timed excretion. Recovery controls and matched non-vesicular fractions are necessary to determine whether a molecular or biological signal is genuinely EV-associated.

Detection of mitochondrial cargo does not establish functional transfer. Convincing evidence should demonstrate cargo association with a characterized EV population, uptake by recipient cells, intracellular delivery to the relevant compartment, and a cargo-dependent biological effect. Donor-specific mtDNA variants, mitochondrially targeted reporters, or complementary labeling of the EV membrane and mitochondrial cargo can track transfer, while recipient-cell oxygen-consumption rate, respiratory reserve, ATP production, membrane potential, respiratory-complex activity, or rescue of mtDNA-deficient cells can assess function. Causality should be tested using donor-cell cargo depletion or modification, EV-depleted and non-vesicular fractions, detergent-disrupted EVs, nuclease or protease controls with membrane permeabilization, and, where feasible, loss-of-function and rescue experiments (Table 1).

Table 1.

Operational workflow for mitoEV-enriched fractions in plasma and urine.

Analytical issue	Recommended approach	Principal limitation or essential control
Plasma pre-analysis	Standardized collection, rapid processing, platelet depletion, and quality assessment	Report anticoagulant, processing time, hemolysis, platelet contamination, storage, and freeze–thaw cycles
Urine pre-analysis	Standardized collection and clearing with documentation of urine concentration	Report creatinine, osmolality, proteinuria, urine volume or flow, uromodulin, and storage
Primary EV enrichment	Size-exclusion chromatography and/or density-gradient separation with an orthogonal method	Assess recovery and co-isolated albumin, lipoproteins, protein aggregates, and non-vesicular DNA
Immunoaffinity enrichment	Optional capture using accessible EV or mitochondrial antigens	TOMM20 or VDAC1 capture alone does not establish mitoEV identity; assess topology and free mitochondrial fragments
EV and mitochondrial identity	Orthogonal EV characterization plus multiple mitochondrial proteins, lipids, or nucleic acids	Avoid classification based on one marker or size range; exclude extracellular mitochondria and fragments
Same-particle confirmation	Nano-flow or imaging flow cytometry, single-particle immunoassay, super-resolution imaging, or immunogold EM	Bulk double positivity does not demonstrate co-localization
Cargo topology	Protease or DNase treatment with and without membrane permeabilization	Distinguish externally associated from protected intravesicular cargo
Quantification and normalization	Report total particles, marker-positive EVs, cargo per particle, and biofluid-specific denominators	Use multiple denominators, recovery controls, and matched non-vesicular fractions
Functional cargo validation	Track donor cargo, demonstrate uptake and compartmental delivery, measure bioenergetic or inflammatory effects, and perform loss/rescue studies	Cargo detection or fluorescence co-localization alone is insufficient

4. Mitochondrial dysfunction as a source of vesicular signaling in CRS

Mitochondrial dysfunction in CRS should be conceptualized as a dynamic and system-level process that extends beyond local bioenergetic failure. In both the heart and kidney, mitochondria act not only as energy-producing organelles but also as central regulators of redox signaling, cell death pathways, and innate immune activation [23–25]. Under pathological conditions, disruption of mitochondrial homeostasis leads to the accumulation of damaged mitochondria, thereby transforming mitochondria into active sources of stress signals rather than passive victims of injury.

In cardiomyocytes, mitochondrial dysfunction is characterized by impaired oxidative phosphorylation, metabolic inflexibility, and excessive mitochondrial reactive oxygen species (mtROS) production [23,24]. These alterations promote mitochondrial DNA (mtDNA) damage, lipid peroxidation, and activation of mitochondrial permeability transition, ultimately contributing to contractile dysfunction and cell death. In parallel, renal tubular epithelial cells—another highly energy-dependent cell type—exhibit early mitochondrial fragmentation, loss of membrane potential, and impaired oxidative metabolism in both acute kidney injury (AKI) and chronic kidney disease (CKD) [26,27]. Persistent mitochondrial injury in the kidney is further associated with fibrosis, inflammation, and progressive loss of renal function.

A critical consequence of mitochondrial dysfunction is the generation and release of mitochondrial damage-associated molecular patterns (mtDAMPs), including mtDNA, oxidized mitochondrial proteins, cardiolipin, and mitochondrial RNA. These molecules can activate cGAS–STING, Toll-like receptor, and inflammasome pathways, linking mitochondrial injury to inflammatory amplification [9,28–30]. Extracellular mtDAMPs, however, are heterogeneous in physical form and may occur as freely circulating or protein-associated molecules, material attached to the EV surface, protected intravesicular cargo, mitochondrial fragments, or extracellular mitochondria. These forms differ in stability, biodistribution, cellular uptake, and receptor accessibility. Free extracellular mtDAMP observations should therefore not be interpreted as mitoEV-specific evidence unless vesicular association and cargo topology have been demonstrated.

EV secretion is one potential route for controlled export of mitochondrial constituents. Under mitochondrial stress or impaired degradation, selected cargo can be packaged into extracellular vesicles [8,9]. When mitophagic or lysosomal handling is insufficient, mitochondrial material may be redirected from intracellular degradation toward extracellular export, although the relative contribution of this route is context-dependent and remains incompletely defined [17].

This proposed mitophagy-to-export shift offers a conceptual bridge between intracellular quality-control failure and systemic signaling. In CRS, sustained mitochondrial stress in the heart or kidney could be externalized through vesicles and influence distant cells; however, this sequence has not been established as a causal cross-organ pathway.

Importantly, this vesicle-mediated release of mitochondrial cargo should not be viewed as a uniform process. The composition and functional impact of mitochondrial cargo-containing EVs are likely to depend on the severity of mitochondrial injury, the efficiency of mitophagy, and the cellular context. These considerations provide the basis for understanding the heterogeneity of mitoEVs, which can either propagate injury or facilitate repair depending on their origin and cargo composition [8,9].

Taken together, mitochondrial dysfunction creates a pool of potentially releasable mitochondrial signals, only a subset of which is vesicle-associated. Defining when, how, and in what physical form these signals leave donor cells is essential before mitoEV-mediated inter-organ communication can be assigned a causal role in CRS.

5. Bidirectional EV/mitoEV-mediated heart-kidney crosstalk in CRS

Direct evidence specifically implicating mitoEVs in CRS is limited. The following sections therefore separate functional data on total EVs from mitoEV-specific findings in related disease models and from theoretical extrapolation to heart–kidney signaling.

5.1. Heart-to-kidney signaling

The strongest heart-to-kidney evidence comes from circulating EVs isolated from patients with CRS, which induced renal epithelial and endothelial injury in a kidney-on-chip system [5]. This constitutes Tier 1 evidence that circulating total EVs from CRS patients can damage renal cells. It does not, however, establish that the vesicles originated from the heart or that mitochondrial cargo mediated the effect.

Tier 2 evidence from myocardial infarction shows that cardiac fibroblast-derived mitochondria-enriched small EVs carrying damaged mitochondrial components can activate NLRP3-dependent inflammation and promote adverse ventricular remodeling [31]. The preparations showed small-EV characteristics and contained mitochondrial signals and proteins, but structurally intact mitochondria were not demonstrated. Accordingly, they are discussed as mitochondria-enriched or mitochondrial cargo-containing small EVs, not as intact-mitochondria-containing EVs. Their ability to injure renal recipient cells or mediate CRS has not been tested; extension of these findings to cardiac mitoEV-mediated renal injury remains a Tier 3 inference.

Systemic inflammatory models provide additional Tier 2 support for vesicular dissemination of mitochondrial stress. In sepsis-induced myocardial dysfunction, microvesicle release was associated with disrupted mitophagic flux and inflammatory amplification [15]. Because the vesicle sources, cargo, and recipient organs differ from clinically defined CRS, these findings support biological plausibility rather than direct heart-to-kidney evidence.

5.2. Kidney-to-heart signaling

CKD provides comparatively strong evidence for kidney-to-heart communication by total EVs. Circulating EVs from patients and experimental CKD models induced cardiomyocyte apoptosis and impaired contractile function, and EV depletion improved cardiac performance [32]. These findings constitute Tier 1 evidence for a pathogenic circulating EV pathway in type 4-like cardiorenal disease, although the validated cargo was predominantly non-mitochondrial.

Renal injury models separately show that EVs can contain or transfer mitochondrial material and alter metabolism or inflammatory responses [8,11]. This constitutes Tier 2 mechanistic evidence that renal stress can generate biologically active mitochondrial cargo-containing EVs. The proposition that kidney-derived mitoEVs account for the cardiotoxic effects observed in CKD remains Tier 3 because organ origin, mitochondrial cargo dependence, and cardiac uptake have not been established in the same causal experiment.

Uremic stress, oxidative injury, and impaired mitochondrial quality control could favor extracellular disposal of mitochondrial material. Whether this response is adaptive, pathogenic, or both will depend on cargo integrity, release kinetics, and recipient-cell context.

Thus, direct evidence supports circulating total EVs as active participants in kidney-to-heart injury, whereas a mitoEV-specific causal mechanism remains unproven.

5.3. Mapping EV and mitoEV evidence to the five CRS subtypes

The five CRS subtypes provide a clinically relevant framework for interpreting EV and mitoEV evidence. Type 1 has direct evidence for renal injury caused by circulating total EVs from CRS patients, but no source-resolved cardiac mitoEV evidence. Type 4 has direct evidence for cardiotoxic total EVs in CKD, whereas the mitochondrial contribution remains indirect. Evidence for mitoEV-mediated signaling in types 2 and 3 is largely extrapolated from chronic heart failure, AKI, or renal mitochondrial-injury models. Type 5 is particularly difficult to resolve because systemic inflammation can simultaneously stimulate EV release from the heart, kidney, endothelium, platelets, and immune cells.

The release kinetics and tissue distribution of mitoEVs may also differ by subtype, but direct longitudinal data are unavailable. Acute types 1 and 3 may involve transient changes in EV abundance and damage-associated cargo, whereas chronic types 2 and 4 may involve sustained exposure to lower-amplitude populations. In type 5, multisource release makes directional attribution especially uncertain. These are testable hypotheses rather than established subtype characteristics.

No reproducible molecular signature currently distinguishes cardiac-derived from renal-derived mitoEVs. Studies differ in biological source, vesicle-size definition, separation method, and mitochondrial readout, precluding valid quantitative comparison of mtDNA copy number, protein oxidation, or lipid composition. Table 2 therefore records organ and cellular source only when supported by the original study, separates observed from proposed cargo, and states the principal evidentiary limitation.

Table 2.

Evidence-graded mapping of EV/mitoEV signaling and orgn-specific features across the five CRS subtypes.

CRS subtype and clinical sequence	Representative evidence/model	Reported or proposed organ/cellular source	Reported or proposed mitochondrial cargo	Recipient organ and functional consequence	Evidence tier and principal limitation
Type 1: acute cardiac dysfunction → AKI	CRS patient-derived circulating EVs injure renal epithelial and endothelial cells; MI mt-sEV studies provide related support	Circulating source unresolved in CRS; cardiac fibroblasts demonstrated in MI; endothelial, platelet, and immune sources are plausible	Not resolved in CRS; damaged mitochondrial proteins/signals detected in MI-derived mt-sEVs	Kidney: endothelial and tubular injury	Tier 1 for total circulating EVs; Tier 2 for MI mitochondrial-EV evidence; cardiac source and mitoEV cargo dependence remain Tier 3
Type 2: chronic cardiac dysfunction → progressive renal impairment	HF-associated EV studies; no direct longitudinal mitoEV-specific heart-to-kidney study	Cardiac, endothelial, platelet, and immune sources proposed	Oxidized or damage-associated mitochondrial cargo proposed but not characterized	Kidney: chronic tubular stress, inflammation, and fibrosis proposed	Predominantly Tier 2–3; no source-resolved causal mitoEV evidence
Type 3: AKI → acute cardiac dysfunction	Renal mitochondrial injury and EV release demonstrated separately in AKI-related models	Renal tubular, endothelial, platelet, and immune sources proposed	Mitochondrial cargo and release kinetics unresolved	Heart: acute inflammation or metabolic dysfunction proposed	Tier 2 for related renal mechanisms; kidney-derived mitoEV cardiotoxicity remains Tier 3
Type 4: CKD → chronic cardiac injury/HF	CKD-associated circulating EVs directly impair cardiomyocytes; validated cargo is mainly non-mitochondrial	Kidney-associated circulating EVs; exact organ/cellular source incompletely resolved	miRNAs established; mitochondrial proteins or mtDNA proposed	Heart: apoptosis, impaired contractility, and remodeling	Tier 1 for total EV cardiotoxicity; mitochondrial cargo evidence is Tier 2 and a mitoEV-specific causal pathway remains Tier 3
Type 5: systemic disorder → simultaneous cardiac and renal injury	Sepsis/systemic inflammatory models show EV release, mitochondrial stress, and inflammatory amplification	Multiple concurrent sources: heart, kidney, endothelium, platelets, and immune cells	mtDNA, mitochondrial proteins, and oxidized lipids reported or proposed across heterogeneous models	Heart and kidney: simultaneous inflammatory and mitochondrial injury	Tier 2 disease-model evidence; source, directionality, and cargo-specific causality are unresolved

Note: ‘Reported cargo’ indicates mitochondrial material detected in the cited study, whereas ‘proposed cargo’ indicates mechanistic inference. Cross-study quantitative comparisons should be interpreted cautiously because biological sources and analytical methods differ.

6. A proposed mitoEV–mitophagy–inflammation axis in CRS

A central unresolved question is whether localized mitochondrial dysfunction can evolve into a systemic, self-amplifying process through vesicular signaling. We propose the mitoEV–mitophagy–inflammation axis as a working hypothesis rather than an established CRS mechanism. Individual links are supported mainly by Tier 2 evidence from related disease models; the complete sequence has not been demonstrated in any CRS subtype.

Under physiological conditions, mitochondrial quality-control pathways remove or segregate damaged material and limit extracellular release of mtDAMPs. Cardiac and renal injury can disrupt mitophagy and mitochondrial quality control [23,26,27]. When lysosomal degradation becomes insufficient, selected mitochondrial material may be redirected toward EV-associated export, as demonstrated in non-CRS models [17].

Several molecular checkpoints may influence cargo fate. PINK1 and Parkin participate in stress-induced MDV formation and lysosomal cargo delivery [33]. OPA1- and SNX9-dependent MDVs can promote selective incorporation of mitochondrial proteins into EVs, whereas Parkin limits this process by directing damaged cargo toward lysosomal degradation[34]. PINK1-dependent packaging of mtDNA into EVs has also been demonstrated in a cancer model [35]. BNIP3, NIX, and FUNDC1 are important mitophagy receptors in hypoxic cardiac and renal stress [36,37], but direct evidence that they regulate extracellular mitoEV release is lacking; they should therefore be regarded as candidate upstream modifiers rather than established mitoEV-biogenesis proteins.

Disease-model evidence also supports links between vesicular mitochondrial cargo and innate immunity. In sepsis-induced myocardial dysfunction, microvesicle release was associated with disrupted mitophagic flux and inflammatory amplification [15]. In recipient cells, EV-associated mtDNA may activate cGAS–STING or endosomal TLR9–MyD88–NF-κB signaling, whereas oxidized mitochondrial proteins or lipids may contribute to NLRP3 inflammasome activation [28–31,35]. The relative importance of these pathways is likely to depend on cargo topology and recipient-cell type, but direct comparisons among cardiac, renal, endothelial, and immune cells in CRS are unavailable.

These observations support a hypothetical feedback sequence in which insufficient mitochondrial quality control favors vesicular export, mitochondrial cargo-containing EVs activate inflammatory responses in recipient cells, and inflammation further impairs mitochondrial function. This model offers a testable explanation for progressive cross-organ injury, but no CRS-specific study has demonstrated the complete causal sequence (Figure 2).

Figure 2.

Evidence-graded heart–kidney crosstalk and the proposed mitoEV–mitophagy–inflammation working model in CRS.

The model is also context-dependent rather than uniformly pathogenic. Stem cell-derived or engineered EVs have shown reparative mitochondrial effects in both renal and cardiac injury models. In AKI, MSC-derived EVs can stabilize mtDNA and restore mitochondrial bioenergetics [11,12], whereas mitochondria-rich EVs from induced pluripotent stem cell-derived cardiomyocytes can transfer functional mitochondria, enhance mitochondrial biogenesis, and improve myocardial energetics and post-infarction cardiac function [38]. Pathogenic versus reparative effects therefore cannot be inferred from mitochondrial cargo alone.

Therapeutic implications of this duality are discussed in Section 7.2. Mechanistic studies should first identify the donor cell, cargo topology, recipient cell, and cargo-dependent functional effect before assigning a specific mitoEV population to the proposed axis.

Validation will require CRS-specific models combining serial and source-resolved mitoEV profiling, direct manipulation of cargo or release pathways, measurement of mitophagic flux in donor and recipient tissues, and simultaneous cardiac and renal outcomes.

The upper panel summarizes bidirectional EV-associated signaling. Circulating total EVs from CRS or CKD settings have direct functional evidence for renal or cardiac injury, whereas the specific contribution, source, and cargo dependence of mitoEVs remain largely indirect or proposed. The lower panel depicts the working sequence of mitochondrial damage, insufficient quality control or mitophagy, vesicular export of mitochondrial cargo, uptake by cardiac or renal recipient cells, activation of cGAS–STING, TLR9, and/or NLRP3 pathways, and secondary mitochondrial dysfunction. The individual links are supported to different extents in related models; the complete feedback sequence has not been established in CRS. Abbreviations: CRS, cardiorenal syndrome; EV, extracellular vesicle; mitoEV, mitochondrial extracellular vesicle; mtROS, mitochondrial reactive oxygen species; mtDNA, mitochondrial DNA.

7. Therapeutic and biomarker implications of mitoEVs in CRS

The recognition that mitochondrial dysfunction can be externalized through EV-mediated pathways raises important translational questions regarding the diagnostic and therapeutic potential of mitoEVs in CRS. While current evidence does not yet establish mitoEVs as definitive drivers of CRS, their ability to reflect mitochondrial status and modulate recipient cell function suggests that they may serve as both biomarkers and therapeutic targets within the broader EV landscape.

7.1. MitoEVs as biomarkers of mitochondrial stress in CRS

One potential application of mitoEV research is biomarker development. Circulating EVs are under investigation in cardiovascular and renal disease [6,7], but total plasma mtDNA is not synonymous with vesicle-associated mtDNA. In plasma, mtDNA may exist as cell-free fragments, protein-associated DNA, EV-surface-associated DNA, protected intraluminal cargo, larger mitochondrial particles, or extracellular mitochondria. Platelet activation, hemolysis, and processing delay can alter these fractions. A candidate mtDNA-positive mitoEV biomarker therefore requires an EV-enriched fraction, orthogonal EV and mitochondrial characterization, DNase analysis before and after membrane permeabilization, matched non-vesicular fractions, and normalization to particle and marker-positive EV counts [39,40].

Urinary EVs can retain molecular features of nephron segments, but their measurement is influenced by hydration, urine flow, osmolality, proteinuria, creatinine, collection time, storage, and uromodulin. No single denominator is sufficient. Results should be reported using complementary measures such as urinary creatinine, particle number, EV-marker abundance, osmolality, timed excretion, and nephron segment-specific markers [22,41]. When urinary EV-associated mtDNA is measured, vesicular and non-vesicular fractions should be separated by size-exclusion chromatography, density-gradient separation, or another orthogonal method, and DNase treatment before and after membrane permeabilization should distinguish free or surface-associated DNA from protected intravesicular mtDNA.

Recent human studies support feasibility but not yet CRS validation. Podocyte-derived urinary large EVs containing mtDNA were associated with relapse in childhood nephrotic syndrome, whereas independently evaluated urinary EV protein signatures have been reported in patients with early renal injury [42,43]. These disease-specific findings demonstrate the potential of the platform but do not establish a mitoEV-specific biomarker for CRS. Disease-model studies have detected mitochondrial cargo in cardiac-derived or circulating EV fractions, but their value as cardiovascular biomarkers remains unvalidated and organ origin and same-particle identity remain incompletely resolved [15,31].

Clinical translation will require prospective, subtype-stratified studies using standardized pre-analytics, the operational identification framework in Section 3, serial cardiac and renal phenotyping, and independent validation cohorts. Biomarker claims should distinguish total EVs, marker-positive EV subpopulations, and cargo that is merely co-isolated or externally associated.

7.2. Therapeutic targeting of mitoEV pathways

From a therapeutic perspective, mitoEV pathways could be targeted by reducing release or uptake of pathological vesicles, neutralizing damaging cargo, or using reparative EVs as delivery platforms. Inhibition of EV formation or release can mitigate injury in selected cardiovascular and renal models, but these studies are not mitoEV-specific and do not establish efficacy in CRS [44].

Stem cell-derived EVs can transfer mitochondrial components, restore bioenergetic function, and reduce inflammation in experimental cardiac and renal injury [11,12]. Engineered EVs may enhance mitochondrial delivery or tissue targeting [12,45], but these findings remain preclinical and should not be extrapolated directly to CRS.

Importantly, the dual nature of mitoEVs necessitates careful consideration in therapeutic design. At present, pathological and reparative mitoEV populations cannot be reliably distinguished by vesicle size or conventional EV-marker profiles alone. A provisional classification should integrate the identity and physiological state of the donor cell, the composition and integrity of the mitochondrial cargo, and the functional response of recipient cells. Operationally, pathological mitoEVs may be defined by their ability to induce inflammatory activation, oxidative stress, or bioenergetic impairment, whereas reparative mitoEVs should demonstrate restoration of mitochondrial respiration, ATP production, or mitochondrial quality control. These categories should remain functional rather than taxonomic until validated source markers and mechanism-linked potency assays become available. Accordingly, indiscriminate suppression of EV release should be avoided because it may inhibit both pathological signaling and adaptive mitochondrial disposal or repair.

Several barriers currently preclude clinical application. Systemically administered EVs may be rapidly cleared or preferentially accumulate in reticuloendothelial organs, limiting selective cardiac or renal delivery. Targeting ligands may alter pharmacokinetics, membrane integrity, immunogenicity, and uptake. Loading of mitochondrial proteins, mtDNA, RNAs, or intact mitochondria remains variable, and cargo activity may be affected by donor-cell conditions, purification, storage, freeze–thaw cycles, or administration. Scalable manufacturing requires control of cell source, culture, purification, sterility, batch consistency, stability, critical quality attributes, and mechanism-linked potency assays. Immunogenicity, complement activation, thrombogenicity, transfer of damaged or mutated mtDNA, and off-target metabolic effects require systematic assessment. Most importantly, no CRS-specific preclinical study has demonstrated that selective inhibition of pathological mitoEVs or delivery of reparative mitoEVs improves both cardiac and renal outcomes. MitoEV-directed therapy should therefore be regarded as a preclinical concept.

7.3. Testable priorities for future CRS research

Three questions should be prioritized over the next five years. First, do the five CRS subtypes exhibit distinct temporal and molecular mitoEV signatures? This requires prospective serial plasma and urine sampling, standardized processing, single-vesicle profiling, and paired cardiac and renal phenotyping. Second, can the organ of origin and causal cross-organ activity of circulating mitoEVs be demonstrated? Lineage-resolved models, donor-specific mitochondrial variants, organ-enriched surface markers, and selective cargo manipulation are needed to test whether heart-derived mitoEVs directly affect the kidney and vice versa. Third, can pathological and reparative mitoEV populations be distinguished and selectively modulated without disrupting physiological EV signaling? This requires validated functional criteria, mechanism-linked potency assays, biodistribution studies, and simultaneous cardiac and renal benefit in CRS-specific models.

8. Conclusion

MitoEVs are biologically plausible but not yet established mediators of CRS. Direct evidence currently supports pathogenic effects of selected total circulating EV populations, whereas mitoEV-specific cross-organ mechanisms are derived mainly from related disease models or theoretical extrapolation. The proposed mitoEV–mitophagy–inflammation axis is therefore a testable working hypothesis, not a definitive causal pathway. Progress will depend on standardized and source-resolved identification, same-particle and cargo-topology validation, functional causality experiments, and prospective studies stratified by CRS subtype. These steps are also prerequisites for determining whether mitoEVs can become reliable biomarkers or selectively targeted therapeutic platforms.

Funding Statement

This study was supported by the Natural Science Foundation of Zhejiang Province of China (LY23H020008), the Sci-Tech Planning Project of Jiaxing (2024AY30010), the Key Medical Disciplines and Specialities Program of Guangzhou (2025–2027), the 2023 Hospital Fund of Guangzhou Panyu Central Hospital (PY-2023-011), the Panyu District Key Medical and Health Project (2024-Z04-004, 2025-Z04-42), and the Major Transverse Research Projects of Jiaxing University (00522195).

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data that support the findings of this study are available from the corresponding author, upon reasonable request.