Nanotherapeutic strategy via ADSC-mitoEVs rescues ischaemic angiogenesis through mitophagy and mitochondrial metabolic reprogramming

Abstract

Ischaemic vascular diseases are critically linked to mitochondrial dysfunction in endothelial cells, which impairs angiogenesis and tissue repair. Although mitochondrial transplantation has emerged as a promising regenerative strategy, its clinical translation remains limited by inefficient delivery and poor retention in target tissues. Here, we demonstrate that mitochondrial-enriched extracellular vesicles derived from adipose-derived stem cells (ADSC-mitoEVs) function as an efficient cell-free nanotherapeutic that restores angiogenic function both in vitro and in a murine model of diabetic hindlimb ischaemia. Mechanistically, ADSC-mitoEV uptake triggers PINK1/Parkin-mediated mitophagy in recipient endothelial cells, a process essential for initiating angiogenesis. Moreover, ADSC-mitoEVs also directly deliver functional mitochondrial proteins, including superoxide dismutase 2 (SOD2), into the endogenous mitochondrial network, which enhances antioxidant activity and improves bioenergetic capacity independently of mitophagy, as demonstrated by reduced reactive oxygen species and elevated ATP production even in PINK1-silenced cells. Our findings establish ADSC-mitoEVs as a versatile cell-free nanotherapeutic that promotes mitochondrial quality control and metabolic reprogramming, offering a potent therapeutic avenue for ischaemic vascular diseases.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s12951-026-04183-x.

Introduction

Ischaemic diseases, encompassing conditions such as peripheral artery disease and coronary artery disease, represent a leading cause of global morbidity and mortality [1]. These pathologies are characterised by a critical imbalance between oxygen supply and metabolic demand within affected tissues [2, 3]. The therapeutic induction of robust and functional angiogenesis remains a paramount, yet persistently elusive, goal in vascular medicine. The ischaemic microenvironment is notoriously pathological, characterised by profound hypoxia, elevated reactive oxygen species (ROS), nutrient deprivation, and sustained inflammation [4]. These factors collectively impose severe bioenergetic stress on resident vascular endothelial cells, compromising their capacity for repair and regeneration.

Central to cellular survival and function, mitochondria are indispensable for energy production, redox homeostasis, and critical signalling processes that govern angiogenesis, including proliferation, migration, and tube formation [5–7]. Compelling evidence now indicates that mitochondrial dysfunction is a fundamental hallmark of ischaemic tissue, directly contributing to cellular apoptosis and the failure of revascularisation [5–7]. Consequently, innovative strategies designed to ameliorate mitochondrial function directly within ischaemic cells represent a compelling and targeted therapeutic approach.

Adipose-derived stem cells (ADSCs) have emerged as a potent source for regenerative therapy, prized for their abundance, accessibility, and pro-angiogenic secretome [8–10]. However, a groundbreaking paradigm shift has occurred with the recognition of horizontal mitochondrial transfer as a fundamental mechanism of ADSC-mediated cytoprotection and repair [11]. Intercellular mitochondrial trafficking can occur via several routes, including the formation of tunnelling nanotubes (TNTs), gap junctions, and cell fusion. While physiologically relevant, these direct cell-contact mechanisms are inherently limited in a therapeutic context by their low efficiency and spatial constraints [11, 12]. In contrast, the transfer of mitochondria via extracellular vesicles (EVs) represents a sophisticated, distance-independent delivery system13. These mitochondria-enriched extracellular vesicles (mitoEVs) are naturally packaged by donor cells and can be efficiently internalised by recipient cells, delivering a full complement of functional proteins, lipids, and mitochondrial DNA (mtDNA) to resuscitate bioenergetic function [13]. This vesicle-mediated mechanism offers distinct advantages for therapeutic application, including enhanced biological stability in circulation, reduced immunogenicity compared with whole cells, and the potential for targeted delivery to sites of injury, thereby bypassing the limitations of direct cell–cell contact. As a nanoscale therapeutic strategy, ADSC-mitoEVs achieve efficient and sustained delivery of mitochondrial cargo to ischaemic tissues. They leverage the intrinsic properties of natural extracellular vesicles, including optimal size for tissue penetration, biological membrane structure for prolonged circulation and bioavailability, and surface molecules for potential cell-specific targeting. This sophisticated convergence of mitochondrial therapy and nanomedicine addresses the translational challenges that have limited existing mitochondrial transfer approaches.

In this study, we isolated and characterised mitoEVs from ADSC-conditioned media using nano-flow cytometry and analysed their protein composition. We then investigated their efficacy in promoting angiogenesis in a murine model of critical hindlimb ischaemia. By utilising integrated proteomic profiling and real-time analysis of cellular metabolism via Seahorse extracellular flux technology, we elucidated the core mechanisms through which ADSC-mitoEVs reprogram cellular energy metabolism and activate potent angiogenic pathways in ischaemic endothelial cells.

Results

Characterisation and function analysis of mitochondrial components in ADSC-EVs

Proteomic profiling identified 390 mitochondrial proteins (8.2% of the total protein content) in ADSC-EVs (Supplementary Fig. 1A and Supplementary Table 1). Gene Ontology (GO) and KEGG pathway analyses highlighted significant enrichment in energy metabolism remodelling, including oxidative phosphorylation, fatty acid β-oxidation, glycolysis, and the regulation of reactive oxygen species (Supplementary Fig. 1B). Functional annotation further revealed roles in mitochondrial targeting, antioxidant activity, proton motive force generation, and electron transport, collectively supporting the potential of ADSC-mitoEVs to modulate cellular bioenergetics (Supplementary Fig. 1C-F).

Characterisation of ADSC-mitoEVs

Both mitoEVs and mito-free EVs from ADSCs exhibited the characteristic cup-shaped morphology revealed by transmission electron microscopy (Fig. 1B). Immunogold electron microscopy confirmed the presence of the extracellular vesicle markers CD63 and TSG101 in both preparations, whilst mitoEVs were additionally enriched with the mitochondrial proteins TOM20 and COX IV (Fig. 1B). Quantitative western blotting revealed that the ratios of mitochondrial proteins to EV-specific markers were markedly elevated in mitoEVs compared with mito-free EVs: the outer mitochondrial membrane protein TOM20 relative to the EV membrane protein CD63 (0.41 ± 0.03 vs. 0.11 ± 0.01), and the inner mitochondrial membrane protein COX IV to the intraluminal EV protein TSG101 (0.61 ± 0.07 vs. 0.08 ± 0.02) (Fig. 1C-E). Consistent with mitochondrial cargo incorporation, mitoEVs showed an increased mean particle size of 218.6 ± 32.5 nm compared with mito-free EVs (105.3 ± 24.1 nm), as determined by nanoparticle tracking analysis (NTA) (Fig. 1F).

Fig. 1.

Isolation and characterisation of ADSC-mitoEVs (A) Schematic illustration of the isolation and sorting process for ADSC-mitoEVs; (B) Representative immunogold electron microscopy of mitoEVs and mito-free EVs for CD63, TOM20, TSG101, and COX IV, colloid gold particles were marked with red triangles, scale bar = 200 nm; (C) Western blot analysis of EV markers (CD63 and TSG101) and mitochondrial markers (TOM20 and COX IV); (D-E) Quantitative analysis of the mitochondrial marker expression relative to canonical EV markers; (F) Nanoparticle size distribution of mitoEVs and mito-free EVs by NTA; (G) Confocal microscopy images showing uptake of ADSC-mitoEVs (red) by mito-DsRed-labelled HUVECs (yellow) under ischaemia (1% O₂, FBS-free) and normoxic conditions (20% O₂, 10% FBS). MitoEVs binding to the mitochondrial network of HUVECs were highlighted by white arrowheads, scale bar = 20 μm; (H) Quantification of the proportion of exogenous mitoEVs bound to the HUVEC mitochondrial network over a period of 4 to 12 h under both ischaemia and normoxic conditions, *p-value < 0.05

Ischaemia accelerates the uptake and mitochondrial binding of ADSC-mitoEVs by HUVECs

We simulated in vivo ischaemia by constructing a cell culture microenvironment defined by hypoxia and serum deprivation. Colocalisation analysis indicated that internalised mitoEVs preferentially bind the endogenous mitochondrial network within 12 h (h) of co-culture, as visualised via confocal microscopy using miRFP670-labelled mitoEVs and DsRed-labelled mitochondria in HUVECs. An increased uptake rate and mitochondrial binding efficiency of mitoEVs were observed in HUVECs cultured under ischaemic conditions, compared with normoxic conditions (Fig. 1G-H).

ADSC-mitoEVs facilitate the survival and angiogenic potential of ischaemic HUVECs

Flow cytometric analysis of apoptosis revealed that ADSC-mitoEVs significantly increased the viability of HUVECs subjected to ischaemia-mimetic conditions (hypoxia and serum deprivation; hereafter referred to as ischaemic HUVECs) in a dose-dependent manner, reaching a peak effect at 40 μg/ml (Fig. 2A-C). Similarly, scratch wound healing and tube formation assays demonstrated markedly enhanced migration and angiogenic capacity in cells treated with mitoEVs compared with untreated controls under ischaemic conditions, with efficacy plateauing at concentrations beyond 40 μg/ml (Fig. 2D-G). Higher doses provided minimal additional benefit to cell survival and angiogenesis. Accordingly, 40 μg/ml was selected as the optimal dose for all subsequent experiments.

Fig. 2.

Cytoprotective and pro-angiogenic effects of ADSC-mitoEVs in vitro. (A-C) Flow cytometric analysis of the effect of ADSC-mitoEVs on ischaemic HUVEC apoptosis. Dose-dependent reduction in apoptosis following ADSC-mitoEV treatment is shown, *p-value < 0.05; (D-E) Representative images and quantification of the scratch motility assay, demonstrating that ADSC-mitoEVs enhanced the migration of ischaemic HUVECs, scale bar = 200 μm, *p-value < 0.05; (F-G) Representative images and quantification of the tube formation assay, showing that ADSC-mitoEVs promoted the tube formation potential of ischaemic HUVECs on Matrigel, scale bar = 500 μm, *p-value < 0.05

Tracking the retention and biodistribution of ADSC-mitoEVs in ischaemic tissue

The fluorescence signal intensity of ADSC-mitoEVs in ischaemic hindlimb tissue was monitored over time. Following administration, the intensity declined progressively but remained detectable for 4 days post-injection (Supplementary Fig. 2A-B). Immunostaining revealed that mitoEVs localised to the subcutaneous fascial vascular network, where they were internalised by endothelial cells (Supplementary Fig. 2C). In muscle tissue, mitoEVs were widely distributed and predominantly sequestered by microvessels and skeletal muscle fibres (Supplementary Fig. 2D).

ADSC-mitoEVs enhance revascularization in vivo

ADSC-mitoEVs significantly improved the angiogenic potential of HUVECs in vivo. Subcutaneous grafts co-implanted with HUVECs and ADSC-mitoEVs exhibited extensive networks of perfused microvessels by day 7, in contrast to control grafts, which showed poor vascularisation (Fig. 3A-G). Furthermore, administration of ADSC-mitoEVs promoted functional recovery in ischaemic hindlimbs, characterised by markedly improved perfusion over three weeks and the prevention of foot necrosis (Fig. 3H-M). Histological analysis revealed increased capillary density (CD31⁺) and enhanced skeletal muscle density and integrity (ACTA1⁺) in ADSC-mitoEV-treated tissues (Fig. 3N-P). These findings demonstrate that ADSC-mitoEVs stimulate robust vascular network formation, enhance tissue perfusion, and protect against ischaemic injury, highlighting their therapeutic potential in driving vascular regeneration and functional recovery.

Fig. 3.

ADSC-mitoEVs promoted revascularisation in vivo. (A) Study design for the in vivo Matrigel plug assay in nude mice; (B) Representative images of new organisms harvested from mice treated with or without ADSC-mitoEVs, scale bar = 0.5 mm; (C) Quantification of surrounding vascular density, *p-value < 0.05; (D) The mass of new organisms, *p-value < 0.05; (E) Representative whole mount immunofluorescence staining for CD31 in Matrigel plugs, scale bar = 200 μm; (F-G) Quantification of vascular density (CD31 positive) and total tube length, *p-value < 0.05; (H) Study design to evaluate the effect of ADSC-mitoEVs on ischaemic hindlimbs in nude mice; (I-J) Representative laser speckle images and quantification indicated improved blood flow in mitoEV-treated hindlimbs, *p-value < 0.05; (K) Statistics of individuals with hindlimb disabilities; (L-M) Representative laser speckle images and quantification indicated improved blood flow in mitoEV-treated feet, *p-value < 0.05; (N-P) Representative whole mount immunofluorescence staining for CD31 and ACTA1 in ischaemic hindlimb muscle tissue demonstrating enhanced revascularisation in the mitoEV-treated tissue, scale bar = 200 μm, *p-value < 0.05

Proteomics indicate improved mitophagy and remodelled energy metabolism by ADSC-mitoEVs

As shown in Supplementary Fig. 3, proteomic profiling of hypoxic HUVECs treated with ADSC-mitoEVs identified 845 significantly upregulated proteins (fold change > 1.25, p < 0.05) associated with autophagy, ubiquitin-mediated protein catabolism, cellular redox homeostasis, mitochondrial ATP synthesis, oxidant detoxification, and inhibition of apoptosis. Key pathways such as oxidative phosphorylation were notably enriched. Additionally, 419 proteins were downregulated (fold change < 0.8, p < 0.05), including those involved in fatty acid β-oxidation. Importantly, GO cellular component analysis revealed that upregulated proteins were enriched in both the “extracellular exosome” and “mitochondrial compartments” terms, supporting successful transfer and uptake of mitoEVs. These findings indicate that ADSC-mitoEVs induce extensive reprogramming of cellular metabolism and enhance cytoprotective responses in endothelial cells under hypoxia, highlighting their role in promoting metabolic adaptation and survival under ischaemic conditions.

ADSC-mitoEVs rescue mitochondrial dysfunction in ischaemic HUVECs

Using JC-1 staining, we observed that ADSC-mitoEVs markedly enhanced mitochondrial membrane potential (MMP) and increased total ATP production in ischaemic HUVECs (Fig. 4 A-C). Furthermore, treatment with ADSC-mitoEVs significantly reduced mitochondrial ROS levels (Fig. 4 D-E). Mitochondrial stress assays demonstrated that ADSC-mitoEVs improved the basal oxygen consumption rate (OCR), mitochondrial ATP production, spare respiratory capacity, and non-mitochondrial ATP generation (Fig. 4 F-J). These results indicate that ADSC-mitoEVs restore mitochondrial bioenergetics, improve redox balance, and boost metabolic flexibility in ischaemic HUVECs, underscoring their therapeutic potential in ameliorating mitochondrial dysfunction.

Fig. 4.

ADSC-mitoEVs rescued the mitochondrial dysfunction in ischaemic HUVECs. (A) Representative JC-1 staining images showing the mitochondrial membrane potential of ischaemic HUVECs treated with or without ADSC-mitoEVs, scale bar = 50 μm; (B) Quantitative analysis of the JC-1 aggregates-to-monomer fluorescence ratio, *p-value < 0.05; (C) Intracellular ATP quantification indicating increased ATP content in mitoEV-treated HUVECs, *p-value < 0.05; (D-E) Flow cytometric analysis of mitoSOX Red stained HUVECs demonstrating reduced mitochondrial superoxide levels in mitoEV-treated HUVECs, *p-value < 0.05; (F-J) Mitochondrial respiration profiles of ischaemic HUVECs demonstrating enhanced basal OCR, ATP-linked respiration, spare respiratory capacity, and non-mitochondrial OCR, *p-value < 0.05

ADSC-mitoEVs promote mitophagy via the PINK1/Parkin signalling pathway

To investigate the impact of ADSC-mitoEVs on mitochondrial quality control under ischaemic conditions, we first employed mt-Keima lentiviral transduction and TEM to monitor mitophagic flux. The results indicated that ADSC-mitoEVs significantly enhanced mitophagy, as evidenced by an increased ratio of lysosomal to cytosolic mt-Keima signal (Fig. 5A-B). Moreover, treatment with mitoEVs preserved mitochondrial morphology under ischaemic stress, as reflected by increased mitochondrial network size (Fig. 5C). TEM demonstrated that ischaemia activated mitophagy, characterised by an increased number of autophagosomes and mitophagosomes. Notably, ADSC-mitoEVs further increased the prevalence of autophagosomes and mitophagosomes under ischaemic conditions (Fig. 5E-G).

Fig. 5.

ADSC-mitoEVs promoted mitophagy of ischaemic HUVECs. (A-C) Representative images of mt-Keima transfected HUVECs demonstrating that ADSC-mitoEVs accelerate mitophagic flux, as indicated by an increased ratio of lysosomal to cytoplasmic signals, and improved mitochondrial network morphology, scale bar = 20 μm, *p-value < 0.05; (D) Absolute quantification by ddPCR showing decreased retention of donor-ADSC-derived mtDNA in HUVECs from 12 to 72 h post-treatment; (E–G) Transmission electron microscopy demonstrated that ischaemia activated mitophagy, evident by an increased number of autophagosomes and mitophagosomes. Treatment with ADSC-mitoEVs further elevated the counts of both autophagosomes and mitophagosomes under ischaemic conditions, autophagosomes were marked with red asterisks scale bar = 5000 nm, *p-value < 0.05

We next evaluated the persistence and incorporation of transferred mitochondria. Droplet digital PCR (ddPCR) using ADSC- and HUVEC-specific mtDNA primers revealed approximately 32,000 copies of ADSC-derived mtDNA in HUVECs at 12 h post-incubation with ADSC-mitoEVs (Supplementary Data 1). This copy number declined to 22,400 and 12,800 at 24 h and 48 h, respectively, and was no longer detectable by 72 h (Fig. 5D, Supplementary Fig. 4, and Supplementary Table 2).

To elucidate the mechanisms underlying augmented mitophagy, we analysed markers of autophagic and mitophagic flux. As shown in Fig. 6, treatment with ADSC-mitoEVs increased the LC3-II/I ratio and decreased p62 expression, confirming enhanced autophagic activity. Western blot analysis further revealed elevated levels of PINK1, phosphorylated ubiquitin, and phosphorylated Parkin, indicating activation of the PINK1/Parkin-mediated mitophagy pathway. In contrast, no significant activation was observed in the BNIP3/NIX pathway.

Fig. 6.

ADSC-mitoEVs promoted mitophagy in ischaemic HUVECs through the PINK1/Parkin pathway. (A, D) Western blot analysis showing that ADSC‑mitoEV treatment increases the LC3‑II/I ratio and reduces p62 expression, indicating enhanced autophagic flux; (C, D) Mitophagy is promoted specifically through the PINK1/Parkin pathway, evident from the upregulation of PINK1, phospho‑ubiquitin, and phospho‑Parkin; (B, D) The BNIP3/NIX pathway was not significantly activated, *p-value < 0.05

ADSC-mitoEVs enhance angiogenesis in PINK1-dependent manner

To investigate whether ADSC-mitoEV-mediated angiogenesis and the restoration of mitochondrial function are PINK1-dependent, we knocked down PINK1 in HUVECs (Supplementary Fig. 5A). PINK1 knockdown abolished the pro-mitophagy and pro-angiogenic effects of ADSC-mitoEVs, as demonstrated by a significant reduction in mitophagic flux (decreased 550/440 nm ratio; Supplementary Fig. 5B-C) alongside impaired tube formation and migration (Fig. 7C-F). However, PINK1 knockdown did not fully abrogate the beneficial effects of ADSC-mitoEVs. Anti-apoptotic activity (Fig. 7A-B), the attenuation of mitochondrial oxidative stress (Fig. 7G-H), and improvements in mitochondrial membrane potential (Fig. 8A-B) and cellular energy metabolism (Fig. 8C-D) were partially retained. Collectively, these data demonstrate that ADSC-mitoEVs operate via a dual mechanism: while PINK1 is indispensable for mitophagic flux and angiogenesis, the restoration of bioenergetic homeostasis and cytoprotection remains partially independent of this signalling axis.

Fig. 7.

ADSC-mitoEVs promote angiogenesis in a PINK1-dependent manner. (A-B) Flow cytometric analysis showing that PINK1 knockdown partially reversed the anti-apoptotic effect of ADSC-mitoEVs, *p-value < 0.05; (C-D) The pro‑migratory effect of ADSC‑mitoEVs was significantly attenuated upon PINK1 knockdown, scale bar = 200 μm, *p-value < 0.05; (E–F) PINK1 knockdown abrogated the pro‑angiogenic activity of ADSC‑mitoEVs, as shown by reduced total tube length, scale bar = 500 μm, *p-value < 0.05; (G-H) The mitochondrial antioxidant effect of ADSC‑mitoEVs was partially attenuated by PINK1 knockdown, *p-value < 0.05

Fig. 8.

PINK1 knockdown attenuates the enhancement of mitochondrial function by ADSC‑mitoEVs. (A-B) Quantification of JC-1 staining showing that the improvement in mitochondrial membrane potential induced by ADSC‑mitoEVs was partially diminished upon PINK1 knockdown, *p-value < 0.05; (C-D) Mitochondrial respiration assay demonstrating that PINK1 knockdown partially reversed the beneficial effects of ADSC‑mitoEVs on mitochondrial energy metabolism, including basal OCR and ATP-linked respiration, *p-value < 0.05

ADSC-mitoEVs facilitate survival of ischaemic HUVECs through SOD2 transfer

To further elucidate the mechanistic basis of ADSC-mitoEV-mediated mitochondrial rescue, we performed Venn analysis to identify mitochondrial proteins enriched in ADSC-mitoEVs and concurrently upregulated in recipient HUVECs (Supplementary Table 2). This approach revealed 43 candidate mitochondrial proteins potentially transferred via mitoEVs (Fig. 9A-B). Subsequent protein-interaction and functional enrichment analyses indicated that these proteins are primarily involved in cellular respiration, mitochondrial organisation, antioxidant responses, and anti-apoptotic pathways. Among these, SOD2 emerged as a central hub capable of modulating these pathways (Fig. 9C).

Fig. 9.

SOD2 as a critical candidate protein for ADSC-mitoEVs to rescue mitochondrial damage. (A) Venn diagram illustrating the intersection of mitochondrial proteins enriched in ADSC-EVs and upregulated proteins in mitoEV-treated HUVECs, identifying 43 candidate proteins; (B) Heatmap showing the expression profile of the candidates in HUVECs with or without mitoEV-treatment; (C) Protein–protein interaction (PPI) and functional enrichment analysis of the candidate proteins using the STRING database. SOD2 functions as a key hub protein involved in multiple biological processes, including cellular respiration, mitochondrion organisation, antioxidant response, cellular biosynthetic process and anti-apoptotic pathways

To experimentally validate the functional significance of SOD2 transfer, we generated SOD2-overexpressing ADSCs and isolated SOD2-enriched mitoEVs (Supplementary Fig. 6A-C). Flag-tagged SOD2 was detected within recipient HUVECs following co-culture, confirming vesicle-mediated delivery (Supplementary Fig. 6D). Western blot analysis further corroborated the increase in SOD2 expression in recipient cells (Supplementary Fig. 6E-F). Treatment with SOD2-enriched mitoEVs significantly reduced apoptosis and mitochondrial ROS levels (Fig. 10A-D), while enhancing mitochondrial membrane potential in PINK-1-silenced ischaemic HUVECs (Fig. 10E-F). Seahorse assays demonstrated improved basal OCR, ATP-linked respiration, spare respiratory capacity, and non-mitochondrial ATP production (Fig. 10 G-K). Accordingly, ADSC‑mitoEVs restore mitochondrial homeostasis and reduce oxidative damage through the intercellular delivery of active mitochondrial proteins, such as SOD2, independent of the recipient’s PINK1/Parkin signalling integrity.

Fig. 10.

SOD2-overexpressed-ADSC-mitoEVs further promoted the survival and mitochondrial homeostasis of ischaemic HUVECs. Compared with control mitoEVs, SOD2-overexpressed-mitoEVs further reduced apoptosis (A-B) and mito-ROS (C-D), while enhancing mitochondrial membrane potential (E–F) and mitochondrial energy metabolism (G-K) in PINK1-silenced cells, scale bar = 50 μm, *p-value < 0.05

Discussion

In this study, we demonstrate that mitochondrial-enriched extracellular vesicles derived from adipose-derived stem cells (ADSC-mitoEVs) function as a potent bioenergetic delivery system capable of rescuing ischaemic endothelial cells through enhanced mitophagy and mitochondrial protein transfer. These vesicles promote angiogenesis, improve mitochondrial function, and facilitate functional recovery in murine models of hindlimb ischaemia, thereby presenting a novel cell-free strategy for ischaemic vascular diseases.

Our experimental strategy was designed to systematically elucidate the functional and mechanistic contributions of ADSC-mitoEVs. Beginning with proteomic characterisation, we identified a significant enrichment of mitochondrial proteins within ADSC-mitoEVs involved in oxidative phosphorylation, redox homeostasis, and anti-apoptotic processes. This initial profiling suggested that mitoEVs might directly supplement dysfunctional mitochondria in recipient cells. However, internalised ADSC-mitoEVs did not integrate into the mitochondrial network of recipient endothelial cells. Instead, they rapidly activated the PINK1/Parkin-mediated mitophagy pathway within 12 h and were fully degraded by 72 h. This activation promoted mitochondrial turnover and quality control, serving as a prerequisite for initiating angiogenesis in ischaemic HUVECs.

Concurrently, mitophagy induction enhanced ROS clearance, alleviating mitochondrial stress [14, 15]. These effects align with mechanisms reported for both free mitochondrial transfer from ADSCs and TNT-mediated mitochondrial delivery[11]. Although the internalisation routes differ (endocytosis for EVs versus direct intercellular transfer for TNTs), both pathways converge on a shared functional outcome: activation of mitophagy and subsequent cytoprotection. We propose that this convergence is mediated by the delivery of mitochondrial components, such as mitochondrial proteins and mtDNA, which are recognised as damage-associated molecular patterns (DAMPs) by the recipient cell’s surveillance system. These patterns trigger PINK1 stabilisation on the outer mitochondrial membrane, Parkin recruitment, ubiquitination of mitochondrial substrates, and ultimately autophagosome formation around the exogenous material [16–18]. The PINK1/Parkin pathway, central to mitochondrial quality control via mitophagy, plays a critical regulatory role in angiogenesis by clearing dysfunctional mitochondria in endothelial cells. By maintaining cellular fitness through limiting ROS, preventing apoptosis, and supporting the necessary metabolic reprogramming for cell migration and proliferation [19–22], this pathway facilitates angiogenesis. Moderate activation of mitophagy is beneficial, as it facilitates the removal of dysfunctional organelles and stimulates mitochondrial biogenesis, leading to increased ATP production [19–22]. Indeed, following mitoEV treatment, we observed increased total mitochondrial content, a well-developed mitochondrial network, and enhanced ATP generation in endothelial cells, indicating improved bioenergetic capacity and cellular fitness under stress conditions.

Notably, unlike free mitochondrial transfer as reported previously [11], ADSC-mitoEVs retained partial functionality, such as antioxidant capacity and improvements in energy metabolism, even in PINK1-silenced HUVECs. We attribute this PINK1-independent effect to the efficient vesicle-mediated delivery of a consortium of functional mitochondrial enzymes, particularly superoxide dismutase 2 (SOD2). Owing to the protective bilayer phospholipid membrane of extracellular vesicles [23, 24], the structural and functional integrity of SOD2 and other oxidative phosphorylation-related proteins is preserved during intercellular transit, enabling them to remain catalytically active upon delivery. Mechanistically, once internalised, SOD2-enriched ADSC-mitoEVs preferentially localise to the mitochondrial compartments of recipient HUVECs, where the encapsulated SOD2 is released and incorporates into the endogenous mitochondrial matrix. The delivered SOD2 dismutates superoxide anions into hydrogen peroxide and oxygen, thereby markedly reducing mitochondrial ROS levels and protecting electron transport chain complexes from oxidative damage, which stabilises mitochondrial membrane potential and enhances ATP synthesis independently of PINK1/Parkin signalling [25–28]. Proteomic analyses further suggest that ADSC-mitoEVs carry additional mitochondrial enzymes, such as ubiquinone oxidoreductase subunits and ATP synthase, which may act synergistically with SOD2. Consequently, ADSC-mitoEVs function as a multi-component therapeutic unit rather than a single-protein carrier to ameliorate bioenergetic deficits and enhance mitochondrial resilience under ischaemic stress. This multi-protein delivery mechanism distinguishes ADSC-mitoEVs from conventional mitochondrial transplantation and underscores their potential as a comprehensive mitochondrial therapy for vascular diseases.

Mitochondrial transplantation has emerged as a promising treatment for conditions such as cancer [29], Parkinson’s disease [30], and myocardial infarction [31]. ADSC-mitoEVs offer distinct advantages over TNT-based mitochondrial transfer. Unlike TNTs, which require direct cell-to-cell contact and are spatially restricted, mitoEVs act as long-range nanocarriers capable of systemic delivery and targeted accumulation in ischaemic tissues. Furthermore, compared with synthetic mitochondrial nanocarriers or liposome‑based mitochondrial delivery systems, ADSC‑mitoEVs retain their native membrane architecture, endogenous signalling molecules, and a natural tropism for injured cells, thereby enhancing biocompatibility and reducing immunogenicity. Beyond their advantages over TNT-based transfer, ADSC-mitoEVs offer a cell-free strategy that combines the advantages of mitochondrial transplantation and conventional EV therapies while overcoming several limitations. Unlike direct mitochondrial transplantation, which suffers from poor stability and short retention times, ADSC-mitoEVs provide a membrane-protected biological vector that enhances protein stability and the delivery efficiency of multiple mitochondrial-derived enzymes. Moreover, compared with heterogeneous conventional MSC-EV preparations, ADSC-mitoEVs represent a defined subpopulation with elevated mitochondrial content, enabling more targeted bioenergetic modulation. Collectively, these attributes establish ADSC-mitoEVs not merely as a vehicle for mitochondrial components, but as an integrated nanotherapeutic system. Their nanoscale dimensions facilitate passive targeting and extravasation into ischaemic tissue, while their endogenous composition ensures biocompatibility and mitigates rapid immune clearance, common hurdles for synthetic nanoparticles. This approach effectively decouples therapeutic efficacy from ADSCS, offering a stable, storable, and precisely dosable nanomedicine for ischaemic repair.

Despite these promising findings, our study has several limitations. First, advanced purification strategies, such as immunoaffinity-based sorting for mitochondrial markers, could yield more homogeneous mitoEV populations once the membrane protein profile of mitoEVs is fully defined. While we identified SOD2 as a central hub protein and demonstrated that its enrichment enhances the therapeutic efficacy of ADSC-mitoEVs, we acknowledge that SOD2 likely acts as a representative of a broader suite of co-delivered mitochondrial antioxidant and metabolic enzymes. Our attempts to generate SOD2-deficient ADSC-mitoEVs revealed that SOD2 is indispensable for ADSC homeostasis; its depletion triggered rapid cellular senescence and mitochondrial collapse, precluding the isolation of functional EVs. Consequently, the beneficial effects of ADSC-mitoEVs should be viewed as a synergistic result of the whole mitochondrial proteome, with SOD2 serving as a critical, but not solitary, mediator of ROS scavenging. Furthermore, while diabetic murine models of hindlimb ischaemia offer useful insights, they do not fully recapitulate the complex pathophysiology of human vascular diseases, particularly in ageing or comorbid settings. Finally, while our histological findings showed increased capillary density (CD31⁺) alongside improved skeletal muscle architecture (ACTA1⁺), further investigation in large animals is essential to determine the long-term impact on muscle force regeneration.

Conclusion

In this study, we demonstrate that ADSC-mitoEVs, a novel cell-free therapeutic platform, ameliorate ischaemic endothelial damage through dual mechanisms: the activation of PINK1/Parkin-mediated mitophagy and the direct transfer of functional mitochondrial proteins. This strategy enhances mitochondrial quality control, redox homeostasis, and energy metabolism reprogramming, offering a promising approach for the treatment of ischaemic vascular diseases.

Materials and methods

Isolation and characteristics of ADSC-mitoEVs

Human lipoaspirates were harvested from the thigh region of healthy female donors (aged 25 years) using water jet-assisted liposuction (Body-Jet System; Human Med AG, Schwerin, Germany). All procedures complied with the principles outlined in the Declaration of Helsinki (1975) and were approved by the Ethics Committee of The Second Affiliated Hospital of Nanchang University (IIT-0–2025-100). Adipose-derived stem cells (ADSCs) were isolated following a previously established protocol [32]. Primary ADSCs were transfected with pLV-mito-miRFP670 lentiviral particles (Han Heng Biotechnology, Shanghai, China) to enable mitochondrial labelling via miRFP670 fluorescence.

Transfected mito-miRFP670-ADSCs (passages 3–5) were cultured in mesenchymal stem cell medium supplemented with 5% extracellular vesicle (EV)-depleted serum at 37 °C under 5% CO₂ and 95% humidified air. The conditioned medium was collected and subjected to sequential centrifugation steps at 800 × g and 3,000 × g at 4 °C to remove cellular debris. The resulting supernatant was concentrated using an ultrafiltration centrifuge tube equipped with a 100 kDa molecular weight cut-off filter (Millipore, MA, USA) by centrifugation at 3,000 × g for 30 min (mins) at 4 °C. The EV suspension was then processed using a CytoFLEX SRT cell sorter (Beckman Coulter, CA, USA) to isolate miRFP670-positive vesicles. Finally, the sorted suspension was filtered through a 0.22 μm membrane and subjected to ultracentrifugation at 110,000 × g for 90 min at 4 °C (Beckman Coulter, CA, USA) to pellet the ADSC-derived mitochondrial extracellular vesicles (ADSC-mitoEVs). (Fig. 1A).

The characterisation of ADSC-mitoEVs was carried out using immunogold transmission electron microscopy (TEM) and nanoparticle tracking analysis (NTA) to determine morphology and particle size distribution. For immunogold TEM, isolated mitoEVs or mito-free EVs were adsorbed onto carbon-coated grids, fixed, and immunolabelled with primary antibodies against CD63, TOM20, COX IV, and TSG101 (Abcam, Cambridge, UK), followed by an 8-nm colloidal gold-conjugated secondary antibody. Samples were visualised using a HITACHI HT7800 (Hitachi, Tokyo, Japan). For NTA, isolated EVs were diluted in sterile, filtered phosphate-buffered saline (PBS) to achieve an optimal concentration of 5 × 10⁷ particles ml⁻¹ for measurement using a ZetaVIEW (Particle Metrix, Munich, Germany). Data were analysed with ZetaView PMX 110 software to calculate the mean, mode, and median particle size, as well as the particle concentration. Each sample was analysed in triplicate. The protein concentration was evaluated using a BCA Protein Assay Kit (Sigma, MA, USA). Western blot analysis for CD63, TOM20, COX IV, TSG101 was conducted for both mitoEVs or mito-free EVs.

Proteomic analysis of ADSC-EVs

Proteomic profiling was employed to characterise the mitochondrial components within ADSC-EVs. Total protein was extracted from isolated ADSC-EVs and analysed by liquid chromatography-tandem mass spectrometry (LC–MS/MS) using a Q Exactive HF mass spectrometer (Thermo Fisher Scientific, MA, USA). The acquired data were processed with MaxQuant software and matched against the UniProt human database. Mitochondrial proteins were identified based on subcellular localisation predictions. Subsequent bioinformatic analyses, including Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment, were performed to delineate significantly enriched pathways associated with the identified mitochondrial proteins.

Cell culture

Human umbilical vein endothelial cells (HUVECs) were acquired from the Chinese Academy of Sciences Cell Bank and maintained in endothelial cell medium (ECM, ScienCell, CA, USA) supplemented with 5% foetal bovine serum (FBS) at 37 °C under normoxic conditions (20% O₂). To simulate ischaemia in vitro, HUVECs were subjected to hypoxia (1% O₂) in serum-free ECM at 37 °C (referred to as ischaemic HUVECs).

MitoEVs uptake

HUVECs were transfected with pLV-mito-DsRed lentiviral particles (Han Heng Biotechnology, Shanghai, China) to achieve mitochondrial labelling with DsRed fluorescence. The transfected mito-DsRed-HUVECs were then co-cultured with ADSC-mitoEVs (10 μg/ml) for 12 h under both normoxic and ischaemic conditions. Cellular uptake of ADSC-mitoEVs was visualised using laser scanning confocal microscopy (LSCM; Carl Zeiss, Oberkochen, Germany). The total fluorescence area of both miRFP670 (donor ADSC-mitoEVs) and DsRed (recipient mitochondria) was quantified in 10 random cells per group. Three independent experiments were performed.

Cell apoptosis assay

Apoptosis was evaluated using an Annexin V/7-AAD Apoptosis Detection Kit (Beyotime Biotechnology, Shanghai, China). After treatment with ADSC-mitoEVs (10, 20, 40 and 80 μg/ml) for 48 h under ischaemic conditions, HUVECs were harvested, washed with PBS, and then stained with Annexin V and 7-AAD for 20 min at 37 °C in the dark. HUVECs cultured under normoxic conditions served as a negative control. Apoptotic rates were analysed using a flow cytometer (BD Biosciences, NJ, USA). Three independent experiments were performed.

Scratch test

A scratch test was performed to evaluate cell migration. HUVECs were seeded in 24-well plates and grown to confluence. A sterile 100 μl pipette tip was used to create a linear scratch. After washing with PBS to remove detached cells, the cells were incubated with ADSC-mitoEVs (10, 20, 40 and 80 μg/ml) in serum-free medium under hypoxia. HUVECs cultured under normoxic conditions served as a positive control. Wound closure was monitored at 0 and 24 h using an inverted microscope (Carl Zeiss, Oberkochen, Germany). The migration rate was quantified using ImageJ software. Three independent experiments were performed.

Tube formation

The angiogenic capability of HUVECs was assessed using a tube formation assay on Matrigel (Corning, NY, USA). Briefly, 96-well plates were coated with Matrigel and polymerised for 30 min at 37 °C. HUVECs were incubated with ADSC-mitoEVs (10, 20, 40 and 80 μg/ml) under hypoxia for 12 h. Tubular structures were imaged under an inverted microscope (Carl Zeiss, Oberkochen, Germany). Total tube length was analysed using the Angiogenesis Analyser tool in ImageJ. Three independent experiments were performed.

In vivo Matrigel plug assay

All animal experiments were approved by the Institutional Animal Care and Use Committee of Nanchang University (NCULAE-20250912001) and complied with the NIH Guide for the Care and Use of Laboratory Animals. For tube formation in vivo, five male BALB/c nude mice (8 weeks old) were anaesthetised by inhalation of isoflurane (induced at 4% and maintained at 2% at a flow rate of 1.5 L/min). 5 × 10⁵ HUVECs with or without ADSC-mitoEVs (40 μg/ml) were suspended in 100 μL Matrigel (Corning, NY, USA). Investigators were blinded to group allocation during procedures and assessments. The Matrigel mixtures were then injected subcutaneously into both axillary regions of the nude mice. After one week, neoformations were harvested and weighed. Surrounding neovascularisation was analysed using Angiogenesis Analyser tool in ImageJ.

Diabetic hindlimb ischaemia model and treatment

Diabetes was induced in 8-week-old male BALB/c nude mice by intraperitoneal injection of streptozotocin (STZ; Sigma, MA, USA) at the dose of 100 mg/kg. Blood glucose levels were monitored from the tail vein using a glucometer (Omron, Kyoto, Japan) on day 14 post-initial injection. Mice with two consecutive fasting (6 h) blood glucose measurements of ≥ 16.7 mmol/L were considered diabetic and included in the experimental cohort. For the hindlimb ischaemia model, ten diabetic BALB/c nude mice were randomly assigned to the mitoEVs group (n = 5) or PBS group (n = 5), with investigators blinded to group allocation during procedures and assessments. Mice were anaesthetised by inhalation of isoflurane (induced at 4% and maintained at 2% at a flow rate of 1.5 L/min), and unilateral hindlimb ischaemia was induced by ligating the femoral artery. ADSC-mitoEVs (40 μg/ml) in 100 μl PBS or an equal volume of PBS as control, were administered via intramuscular injection at three sites in the ischaemic muscle immediately after surgery. Blood perfusion was monitored on days 0, 7, 14 and 21 using a Laser Speckle Contrast Imaging system (Perimed, Stockholm, Sweden). For biodistribution analysis, mice were imaged at 24, 48, 72, and 96 h post-injection using a PerkinElmer IVIS Lumina LT Series III (Revvity, MA, USA) with excitation at 640 nm and emission at 670 nm. Fluorescence signal intensity (radiance, p/s/cm²/sr) at the injection site was quantified using Living Image software with background signal from the contralateral limb subtracted. At 24 h after injection, subcutaneous fascia rich in vascular beds as well as skeletal muscles were collected from the thighs, and whole-mount staining was performed to observe the uptake of mitoEVs by vascular endothelial cells. The thigh muscle tissue was harvested on the 21st day after surgery and used for subsequent experiments.

Whole mount staining

Neoformation and thigh muscle tissues from nude mice were fixed in 4% paraformaldehyde for 24 h, followed by dehydration in 30% sucrose for 72 h. The samples were then embedded and cryosectioned into 150 μm-thick slices. For immunofluorescence staining, neoformation sections were incubated with an anti-CD31 primary antibody (R&D Systems, MN, USA). Thigh muscle sections were co-stained with anti-CD31 and anti-alpha actin 1 primary antibodies (R&D Systems and Sigma, respectively). Neovascular density and tube length were visualised using laser scanning confocal microscopy (LSCM; Carl Zeiss, Oberkochen, Germany) and quantified using the Angiogenesis Analyser tool in ImageJ software. Muscle density was assessed using ImageJ.

Proteomic analysis of ADSC-mitoEVs regulating ischaemic vascular endothelial cells

HUVECs were treated with ADSC-mitoEVs (40 μg/ml) under ischaemic conditions for 24 h, followed by total protein extraction. Liquid chromatography-tandem mass spectrometry (LC–MS/MS) was performed on a Q Exactive HF mass spectrometer (Thermo Fisher Scientific, MA, USA). Data were analysed using MaxQuant software and searched against the UniProt human database. Bioinformatic analysis, including GO and KEGG enrichment, was conducted to identify significantly altered pathways. Three independent experiments were performed.

Mitochondrial membrane potential assay

Mitochondrial membrane potential was measured using the JC-1 Kit (Beyotime, Shanghai, China). After treatment, HUVECs were incubated with JC-1 staining solution for 20 min at 37 °C, and visualised using laser scanning confocal microscopy (LSCM; Carl Zeiss, Oberkochen, Germany). The ratio of red (aggregates) to green (monomers) fluorescence was calculated to assess MMP. Three independent experiments were performed.

Mitochondrial respiration and oxidative stress assay

Mitochondrial respiration was assessed using a Seahorse XFe24 Analyser (Agilent, CA, USA). HUVECs were seeded in XF24 plates and treated with ADSC-mitoEVs (40 μg/ml) under ischaemic conditions. The oxygen consumption rate (OCR) was measured under basal conditions and in response to oligomycin, FCCP and rotenone/antimycin A as per the manufacturer’s recommendation. For mitochondrial superoxide detection, cells were stained with mitoSOX Red (Invitrogen, CA, USA) and analysed by flow cytometry (BD Biosciences, NJ, USA). Three independent experiments were performed.

ATP measurement

Intracellular ATP levels were quantified using an ATP Assay Kit (Abcam, Cambridge, UK) based on the luciferase-luciferin reaction. Treated HUVECs were lysed and luminescence was measured using a microplate reader (ShanpuBiotech, Shanghai, China). ATP concentrations were normalised to total protein content. Three independent experiments were performed.

Mitophagy flow assay

HUVECs were transfected with mt-Keima lentivirus (Han Heng Biotechnology, Shanghai, China) following the manufacturer’s protocol. To assess mitophagic flux, cells were pre-treated with Bafilomycin A1 (50 nM) for 16 h to block autophagosome-lysosome fusion. Following treatment with ADSC-mitoEVs under ischaemic conditions, mitophagy was assessed using laser scanning confocal microscopy (LSCM; Carl Zeiss, Oberkochen, Germany) and TEM.

Mitophagy activity was quantified based on the ratio of lysosomal (550 nm excitation) to cytoplasmic (440 nm excitation) mt-Keima signals in 30 random cells per group per time point. Furthermore, autophagosomes and engulfed mitochondria were quantified by examining 30 fields of view at 5000 × magnification for each group and time point under TEM (Hitachi, Tokyo, Japan). Additionally, key proteins associated with mitophagy pathways, including PINK1, phospho-ubiquitin, phospho-Parkin (PINK1/Parkin pathway), BNIP3/NIX, LC3-II/I, and P62, were analysed by western blot. Three independent experiments were performed.

Exogenous mtDNA assay

To evaluate the persistence of transferred mitochondria in recipient HUVECs over time, we employed a mitochondrial DNA (mtDNA) genotyping strategy based on the detection of heteroplasmy. mtDNA was isolated from ADSCs and HUVECs using a Mitochondrial DNA Isolation Kit (MCE, NJ, USA). Regions with sequence divergence between the two cell types were identified by Sanger sequencing. Specific primers were designed to target these cell-type-specific alleles, and their amplification efficiency was validated by droplet digital PCR (ddPCR) on a QuantStudio Absolute Q system (Thermo Fisher Scientific, MA, USA). Using these selective primers, we quantified the absolute copy number of ADSC-derived mtDNA in recipient HUVECs at 12, 24, 48, and 72 h post-treatment. Three independent experiments were performed.

Lentiviral-mediated PINK1 knockdown

Short hairpin RNA (shRNA) lentiviral particles targeting human PINK1, along with negative control shRNA constructs, were obtained from Han Heng Biotechnology (Shanghai, China). HUVECs were transduced with lentiviral particles in the presence of polybrene. Following a 72 h incubation, transduced cells were selected using puromycin. Silencing efficiency was confirmed by qRT–PCR analysis of mRNA extracted from selected cells. Three independent experiments were performed.

Generation of SOD2-overexpressed ADSC-mitoEVs

To determine whether ADSC-mitoEVs confer mitochondrial protection in recipient cells via delivery of mitochondrial proteins, we identified overlapping mitochondrial proteins between the mitoEVs and upregulated mitochondrial proteins in HUVECs treated with mitoEVs using Venn analysis. This set was defined as candidate mitochondrial factors. Protein–protein interaction and functional enrichment analyses were performed using STRING (https://cn.string-db.org/) to predict the biological roles of these candidates, leading to the identification of SOD2 as a central hub protein.

Lentiviral particles encoding human flag-tagged SOD2 and the corresponding empty vector controls were obtained from Han Heng Biotechnology (Shanghai, China). ADSCs were transduced with the lentiviral particles in the presence of polybrene. After 72 h, stable cell lines were selected using puromycin (1 µg/ml). Overexpression efficiency was confirmed at the mRNA level by qRT‑PCR and at the protein level by western blot and immunofluorescence staining, using antibodies against SOD2 (Santa Cruz Biotechnology, Dallas, USA) and Flag (Proteintech, Wuhan, China).

SOD2-overexpressing ADSCs were cultured, and the conditioned medium was collected for the isolation of SOD2‑overexpressing ADSC‑mitoEVs (SOD2‑mitoEVs). The uptake of SOD2‑mitoEVs by HUVECs was analysed via Flag immunofluorescence staining. Additionally, SOD2 expression in HUVECs that had taken up SOD2‑mitoEVs was further evaluated by western blot. Three independent experiments were performed.

Statistical analyses

Data are presented as mean ± standard deviation, derived from three independent in vitro experiments or five independent in vivo experiments. Statistical significance between groups was analysed via unpaired Student’s t-tests or one-way analysis of variance (ANOVA) followed by Tukey's honestly significant difference (HSD) test (GraphPad Prism 10, CA, USA). Significance was defined as P < 0.05. A detailed evaluation of all samples was performed by three of the authors blinded to sample identity.

Author contributions

Yuan-zheng Zhu contributed to the study design, isolation and characterisation of ADSC-mitoEVs, cell culture and functional assay *in vitro*, proteomics assay and manuscript writing. Min-chen Zhang contributed to Seahorse assay and ATP detection. Xue-er Li contributed to the heterogeneity analysis of mtDNA and generation of PINK1 silenced and SOD2 overexpressed cells. Xing-hong Zeng contributed to the animal experiments. Xue-ting Gong and Yu-zi Wu contributed to western blotting. Ze-jun Dong, Shu Wu, and Xue-fei Liu contributed to statistical analysis. Abdul Haseeb Khan reviewed and edited the manuscript. Yang-yan Yi contributed to the conception, design, financial support and final approval of the manuscript. All authors read and approved the final manuscript.

Funding

This work was supported by the National Natural Science Foundation of China [Grant number: 82460448]; Natural Science Foundation of Jiangxi Province [Grant number: 20242BAB26140 and 20252BAC240585]; Jiangxi Province Key laboratory of Precision Cell Therapy [Grant number: 2024SSY06241].

Data availability

All data generated and analysed during this research are included in this published article.

Declarations

Ethics approval and consent to participate

All animal experiments were approved by the Institutional Animal Care and Use Committee of Nanchang University (NCULAE-20250912001) and complied with the NIH Guide for the Care and Use of Laboratory Animals. Human lipoaspirates complied with the principles outlined in the Declaration of Helsinki (1975) and were approved by the Ethics Committee of The Second Affiliated Hospital of Nanchang University (IIT-0–2025-100).

Consent for publication

All authors agree for publication.

Competing interests

The authors declare no competing interests.