Mitochondria transplantation reduces intraocular pressure by improving mitochondrial function and remodeling trabecular meshwork

Abstract

Background

Mitochondrial transplantation represents a promising therapeutic strategy for diseases characterized by mitochondrial dysfunction, with mitochondrial dysfunction in trabecular meshwork (TM) recognized as a key pathogenic mechanism in primary open-angle glaucoma (POAG) where impairment drives sustained intraocular pressure (IOP) elevation.

Objective

This study aimed to develop a novel therapeutic approach through direct mitochondria transplantation to replenish mitochondrial quantity in TM cells.

Methods

The healthy mitochondria were isolated from mouse liver tissues, and uptake by TM cells and their ability to alleviate oxidative stress were evaluated under varying oxidative stress conditions, using both 2D monolayers and 3D spheroid models. The underlying mechanisms were explored through proteomic analysis and western blot. The intraocular distribution and therapeutic efficacy, including IOP -lowering effects and TM structural improvement, were assessed in a chronic OHT animal model.

Results

TM cells actively internalized exogenous mitochondria, which were subsequently enriched within the cellular mitochondrial pool. Moderate oxidative damage enhanced this mitochondrial uptake. At both 2D and 3D cellular levels, exogenous mitochondria effectively reduced oxidative stress and enhanced cell viability. This cytoprotective effect was mediated primarily by suppressing excessive mitophagy, thereby restoring autophagic homeostasis and increasing cell survival. In vivo tracking revealed a preferential accumulation of exogenous mitochondria in the TM region, a phenomenon that was more pronounced in the chronic OHT model. A single intracameral injection of mitochondria produced a potent, sustained, and stable reduction in IOP. This therapeutic effect surpassed that of two first-line clinical hypotensive agents (Timolol maleat and Latanoprost eye drops) and was well-tolerated.

Conclusion

This research provides the first systematic evidence establishing mitochondrial transplantation as an efficacious anti-glaucoma therapy, proposing innovative IOP-lowering strategies for glaucoma treatment.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12967-026-07964-y.

Introduction

Glaucoma, a heterogeneous group of optic neuropathies characterized by progressive neurodegeneration of retinal ganglion cells and corresponding visual field loss, remains the leading cause of irreversible blindness worldwide [1, 2]. Primary open-angle glaucoma (POAG), the most common subtype, is projected to affect over 100 million individuals globally by 2040, representing 71.3% of all glaucoma cases [3, 4]. The disease’s insidious onset, phenotypic variability and irreversible progression underscore its status as a major public health challenge, imposing severe socioeconomic burdens on healthcare systems. While intraocular pressure (IOP) reduction remains the cornerstone of POAG management, current therapeutic strategies face significant limitations. Pharmacological monotherapy often proves inadequate for sustained IOP control, necessitating complex combination regimens that are frequently compromised by adverse effects, poor long-term adherence, and suboptimal patient-reported outcomes [5, 6]. Surgical interventions, though valuable, exhibit marked interindividual variability in efficacy, with many cases demonstrating unsatisfactory long-term IOP stability [7, 8]. These critical gaps highlight the urgent need for novel, targeted approaches to optimize IOP modulation and neuroprotection in glaucoma therapy.

The pathogenesis of POAG is complex, involving interactions among multiple pathophysiological processes, including ischemia and hypoxia, extracellular matrix remodeling, genetic mechanisms, immunological dysregulation, and mitochondrial dysfunction, which have not yet been fully elucidated [9–12]. Among these, mitochondrial dysfunction in trabecular meshwork (TM) cells has been recognized as one of the key pathogenic mechanisms in POAG, with studies linking POAG to mitochondrial DNA (mtDNA) mutations, oxidative damage, and bioenergetic failure in trabecular meshwork (TM) cells [13–16]. Histopathological analyses of POAG-affected TM tissues demonstrate ultrastructural mitochondrial defects-including edema, inner membrane rupture, and cristae disorganization-alongside TM cell detachment and lysis [17–19]. These morphological aberrations correlate with functional deficits: POAG-derived TM cells exhibit reduced ATP synthesis, loss of mitochondrial membrane potential (ΔΨm), and exacerbated mtDNA damage [20, 21]. Mechanistically, elevated IOP induces mitochondrial oxidative stress in TM cells, characterized by excessive ROS production and impaired antioxidant defenses. This imbalance propagates a self-perpetuating cycle of mitochondrial depletion, TM cell apoptosis, and pathological extracellular matrix remodeling, collectively obstructing aqueous humor outflow and exacerbating IOP elevation [16, 18, 20, 21]. Therapeutic strategies targeting mitochondrial homeostasis-such as TM bioenergetic restoration, structural preservation, and aqueous drainage pathway reconstruction-thus represent promising avenues for mitigating POAG progression.

Mitochondria transplantation, first demonstrated by McCully and colleagues, is an emerging therapeutic strategy involving the isolation of functional mitochondria from a patient’s healthy autologous tissues and their subsequent delivery to ischemic or damaged organs to restore cellular homeostasis [22]. This approach capitalizes on the inherent biological properties of mitochondria-including replication, dynamic fission/fusion, mitophagy, and intercellular transfer-to facilitate the selective replacement of dysfunctional organelles [23]. Mitochondria supplementation has shown therapeutic potential across diverse pathologies, ranging from neurological disorders and ischemic conditions to oncological diseases [24–28]. Many studies reveal that transplanted mitochondria enhance recipient cell viability by restoring electron transport chain activity, reducing oxidative stress through ROS scavenging, and attenuating apoptosis via improved bioenergetic capacity [29–31]. Jiang et al. reported in a corneal injury model that mitochondrial delivery from stem cells to corneal epithelial cells reduced oxidative stress and enhanced corneal repair, supporting the feasibility of mitochondria transplantation for treating ocular diseases [32, 33]. Critically, in the context of TM pathology, conventional pharmacological strategies aiming to rescue endogenous mitochondrial function often prove inadequate due to the irreversible structural and functional damage characteristic of POAG-associated mitochondrial dysfunction. In contrast, mitochondrial transplantation offers a paradigm-shifting alternative: by delivering intact mitochondrial units with normal genomic and metabolic competence, this approach may directly compensate for mtDNA defects and restore TM cellular function. Such targeted mitochondrial reconstitution could fundamentally address the root cause of IOP dysregulation, positioning it as a transformative strategy for glaucoma management.

Based on these findings, we established novel evidence that POAG was a mitochondrial disorder through studies using a dexamethasone-induced ocular hypertension (DEX-induced OHT) model that recapitulated characteristic TM pathology. Our demonstration that anterior chamber injection of mitochondria could restore TM tissue integrity and regulate IOP, revealed three key therapeutic effects: sustained improvement in TM mitochondrial function, enhanced aqueous humor outflow capacity, and rapid, durable IOP lowing. Mechanistically, this intervention reestablished autophagic homeostasis in TM cells by normalizing mitophagy levels, which reduced intracellular oxidative stress while increasing ATP production and stimulating TM cell viability. These cellular improvements collectively drove structural remodeling of the TM tissue, with the additional benefit of promoting biogenesis of healthy mitochondria, ultimately restoring physiological aqueous humor outflow and achieving sustained IOP control. Our work not only established mitochondrial functional status as a critical determinant of TM health in glaucoma but also positioned mitochondrial transplantation as a transformative treatment strategy for mitochondrial dysfunction-associated glaucoma, offering both rapid therapeutic effects and long-term tissue restoration (Fig. 1).

Fig. 1.

Schematic illustration showing the mechanism of exogenous mitochondrial transplantation via anterior chamber injection in DEX-induced OHT model. The DEX-induced OHT model exhibits mitochondrial structural disruption in TM cells, triggering excessive autophagy and abnormal energy supply. This cascade induces oxidative stress damage, aberrant TM cells apoptosis, and reduced cell density, ultimately impairing aqueous humor outflow and elevating IOP. Following anterior chamber injection of exogenous mitochondria, the intervention effectively augments mitochondrial mass within TM cells and stimulates more healthy mitochondrial biogenesis. Concurrently, it restores mitophagy homeostasis, enhances cellular energy levels, and reduces intracellular ROS. These processes collectively enhanced TM cells viability, normalized TM cells population, and remodeled TM architecture, thereby restoring aqueous outflow function and achieving IOP reduction

Results

Oxidative preconditioning potentiated exogenous mitochondria uptake in TM cells

Exogenous mitochondria were successfully isolated from multiple tissue sources, including hepatic, cardiac, muscle tissues, as well as stem cells and tumor cells [34]. Hepatocytes are among the cell types with the highest mitochondrial density in the body, and their mitochondria exhibit particularly active oxidative phosphorylation and energy metabolism functions. This high-energy characteristic may more effectively repair the energy metabolism deficits in TM cells. Liver tissue is easily accessible, and differential centrifugation allows for the isolation of highly pure and active mitochondria. Compared to other tissues, liver-derived mitochondria offer more stable and reliable yield and quality. Previous studies have shown that liver-derived mitochondria demonstrate good therapeutic effects in disease models such as myocardial ischemia [24, 25, 28]. Drawing on these successful experiences, we applied this approach for the first time to the treatment of OHT. In our study, functional mitochondria were efficiently purified from liver tissue using optimized differential centrifugation at 4℃, with an extraction rate of approximately 38.611 ± 2.11% (Supplementary Fig. 1). Dynamic light scattering characterization confirmed the mitochondrial preparations exhibited a homogeneous size distribution with a predominant peak at 650 nm (Fig. 2B). TEM imaging revealed that the isolated mitochondria retained their native ultrastructure, with well-defined inner and outer membranes remaining intact, which proved crucial for their subsequent functional applications (Fig. 2C). The inner membrane displayed characteristic cristae folding patterns without evidence of pathological swelling or rupture. These ultrastructural observations confirmed the preservation of organelle architecture essential for mitochondrial function. Following prolonged ultrasonic treatment, the mitochondria exhibited severe edema, significant swelling, and complete disintegration of their membrane architecture (Supplementary Fig. 2).

Fig. 2.

Extraction, characterization and cellular uptake of exogenous mitochondria in TM cells. (A) Schematic diagram of mitochondrial isolation from mouse liver; (B) Particle size distribution of isolated mitochondria; (C) TEM of isolated mitochondria (X15k, scale bar = 500 nm.); (D) Mitochondrial quality confirmation by MitoTracker staining under confocal laser scanning microscopy (Scale bar = 20 μm); (E) Mitochondrial membrane potential (ΔΨm) detection using JC-1 assay kit. CCCP, a kit-provided positive control, collapses ΔΨm. Red fluorescence indicates JC-1 aggregates, means high ΔΨm, while green represents JC-1 monomers, means low ΔΨm (Scale bar = 50 μm). (F) Fluorescence emission spectra of JC-1-stained mitochondria. The red peak indicates JC-1 aggregates, whereas the green peak represents JC-1 monomers. (Gi-Gii) Schematic of exogenous mitochondrial uptake in TM cells. (Gi) Mitochondria pre-labeled with MitoTracker Red were co-cultured with CFDA-SE (green)-labeled TM cells (mitochondrial protein concentration: 2 µg proteins/mL). (Gii) Uptake assay in oxidatively damaged TM cells pre-treated with H₂O₂. ‌(Hi)‌ The qualitative analysis of mitochondrial uptake efficiency in TM cells at different time by CLSM.‌ (Hii) Time-fluorescence curve of mitochondrial uptake by TM cells plotted based on Figure Hi. ‌(Hiii)‌ Qualitative analysis of mitochondrial uptake efficiency in TM cells under different damage conditions by CLSM. Blue, nuclei; green, TM cells; red, exogenous mitochondria (Scale bar = 10 μm). (Ki) Flow cytometry quantification of the effect of different damage conditions of TM cells on mitochondrial uptake efficiency. (Kii) Mean red fluorescence intensity in TM cells (n = 3). (Li) TEM observation of mitochondria in TM cells. Green arrows, healthy mitochondria, yellow arrows, damaged mitochondria (Scale bar = 500 nm). (Lii) Counting of healthy mitochondria in TM cells (n = 3). Data are expressed as mean ± standard deviation. Statistical significance was calculated by ordinary one-way ANOVA

We next assessed mitochondrial membrane potential (ΔΨm) using two complementary approaches. First, MitoTracker Deep Red 633 fluorescence intensity analysis confirmed that the isolated mitochondria retained strong ΔΨm-dependent dye accumulation (Fig. 2D). These findings were corroborated by JC-1 staining. JC-1 is a cationic lipophilic fluorescent dye that exists in two forms: monomer and aggregate. In healthy mitochondria with polarized membrane potential, JC-1 accumulates in mitochondria to form aggregates emitting red fluorescence. When mitochondrial membrane potential is depolarized, JC-1 releases from mitochondria and exists as monomers emitting green fluorescence. Thus, the ratio of red-to-green fluorescence intensity reflects mitochondrial membrane potential status [35]. After staining the extracted mitochondria suspension with JC-1, healthy mitochondria showed intense red fluorescence with minimal green fluorescence (Fig. 2E). Treatment with CCCP (a mitochondrial uncoupling agent that dissipates membrane potential) significantly increased green fluorescence, demonstrating substantially lower membrane potential compared to untreated mitochondria. Spectral scanning analysis of JC-1 further validated these observations (Fig. 2F). Collectively, these data demonstrated that the isolated mitochondria maintained both structural integrity and functional membrane potential.

We subsequently established a classical oxidative stress injury model in TM cells using H₂O₂ induction. To determine the optimal treatment conditions, we first assessed concentration-dependent cellular responses through MTT viability assays. The MTT assays revealed a concentration-dependent cytotoxic effect of H₂O₂ on TM cells. While cell viability remained > 85% at concentrations below 1.0 mM, exposure to 2.0 mM and 3.0 mM H₂O₂ significantly reduced viability to 75.8 ± 3.08% and 50.15 ± 4.91%, respectively (p < 0.001 vs. untreated controls; Supplementary Fig. 3). Based on these MTT findings, we selected 1.0 mM, 2.0 mM and 3.0 mM H₂O₂ to establish corresponding low-, moderate- and high-level oxidative damage models in TM cells.

To characterize the cellular internalization dynamics of exogenous mitochondria, we systematically examined time-dependent uptake patterns. Mitochondria were firstly labeled with MitoTracker Deep Red 633 to enable tracking and directly added to TM cells (Fig. 2Gi). As shown in Fig. 2Hi-Hii, the labeled mitochondria exhibited intense red fluorescence confirming structural integrity and preserved membrane potential in the isolated mitochondria. Furthermore, it can be seen from the Fig. 2Hi-Hii that the uptake of mitochondria showed time-dependent. After 1 h of incubation, faint red fluorescence was observed within TM cells. The uptake intensity increased with prolonged incubation time, showing significantly enhanced intracellular red fluorescence at 12 h, with predominant perinuclear localization.When incubation was extended to 24 h, the fluorescence intensity remained comparable to that at 12 h, suggesting saturation of uptake by 12 h. These findings established that simple co-incubation enabled efficient mitochondrial internalization, with kinetics reaching completion within 12 h. We therefore selected this timepoint as the optimal incubation time for all subsequent experiments. Additionally, we examined the uptake of MitoTracker Deep Red 633-stained mitochondria by TM cells, observing minimal change in intracellular fluorescence intensity between 24 and 48hs, which suggested no significant dye leakage occurred within this timeframe (Supplementary Fig. 4). Further, we labeled endogenous and exogenous mitochondria with green and red fluorescent markers, respectively, and observed their colocalization (Supplementary Fig. 5). The results demonstrated that exogenous mitochondria colocalized with the endogenous mitochondrial pool, providing a foundation for the functional contribution of the transplanted mitochondria.

To examine oxidative stress-dependent mitochondrial uptake, we established graded injury models in TM cells using 1.0 mM (mild), 2.0 mM (moderate) and 3.0 mM (severe) H₂O₂ pretreatment. Then TM cells were labeled with green fluorescence, while exogenous mitochondria were labeled with MitoTracker Deep Red 633 (Fig. 2Gii). CLSM images revealed striking stress-dependent uptake patterns (Fig. 2Hiii). When cells treated with 1.0 and 2.0 mM H₂O₂ showed no significant impairment in their capacity to uptake exogenous mitochondria, as evidenced by strong intracellular red mitochondrial fluorescence. Notably, compared to the control group, we were surprised to find that cells subjected to 2.0 mM H₂O₂ oxidative stress exhibited even stronger overall red fluorescence intensity than control groups. In contrast, cells treated with 3.0 mM H₂O₂ displayed the weakest intracellular fluorescence. These findings suggested that moderate oxidative stress enhances cellular uptake of exogenous mitochondria, whereas excessive oxidative damage diminishes this capability. Flow cytometric quantification confirmed this biphasic response (Fig. 2Ii-Iii), demonstrating peak mean fluorescence intensity following 2.0 mM H₂O₂ moderate oxidative stress.

Ultrastructural analysis by transmission electron microscopy (TEM) revealed progressive oxidative damage and subsequent therapeutic rescue in TM cells (Fig. 2Ji). Control cells revealed mitochondria with intact double membranes displaying clearly defined inner and outer boundaries, well-organized cristae exhibiting tightly packed lamellar and uniform size distribution without pathological swelling or fragmentation. TEM analysis revealed progressive mitochondrial swelling with increasing H₂O₂ exposure, characterized by disrupted cristae and matrix vacuolization. Concurrently, the population of intact, healthy mitochondria decreased significantly (p < 0.01), exhibiting dose-dependent fragmentation (Fig. 2Jii). Exogenous mitochondria supplementation differentially modulated these effects, both 1.0 and 2.0 mM H₂O₂ groups exhibited significant mitochondrial population recovery (1.0mM vs. 1.0mM+mito = 3.67 ± 1.15 vs. 4.67 ± 0.58, p = 0.184; 2.0mM vs. 2.0mM+mito = 2.0 ± 0.01 vs. 9.33 ± 0.58, p < 0.001), with the 2.0 mM H₂O₂ group demonstrating superior therapeutic outcomes through maximal mitochondrial density and preservation of structurally intact organelles, thereby establishing a robust foundation for functional restoration. Conversely, the 3.0 mM H₂O₂ group showed minimal therapeutic response, with persistent mitochondrial degeneration indicating irreversible oxidative damage despite marginal increases in healthy mitochondria. These results established that moderate oxidative preconditioning (2.0 mM H₂O₂) creates an ideal microenvironment for mitochondrial integration and functional restoration, while severe stress exceeds cellular repair capacity.

Collectively, these findings demonstrated that the isolated exogenous mitochondria maintained structural integrity and functional competence, enabling efficient uptake by TM cells. The internalization process reached optimal efficiency following 12 h of co-incubation, with cells subjected to moderate oxidative stress (2.0 mM H₂O₂) exhibiting the most robust mitochondrial incorporation capacity.

Mitochondrial transplantation restored TM cell bioenergetics and viability

JC-1 analysis revealed that exogenous mitochondria transplantation restored ΔΨm in oxidatively stressed TM cells. Untreated controls exhibited predominant red fluorescence (JC-1 aggregates), indicating intact ΔΨm (Fig. 3Ai). H₂O₂ treatment (2.0–3.0 mM) induced dose-dependent fluorescence shift to green (monomers), confirming ΔΨm loss. Mitochondria supplementation significantly reversed this effect, with both red and green fluorescence intensities particularly in 2.0 mM H₂O₂-treated cells recovering to levels comparable to untreated controls. CCCP-treated positive controls exhibited near-complete green monomer fluorescence, confirming maximal ΔΨm dissipation. Subsequently, we used ImageJ to analysis the fluorescence intensity and calculate the relative ratio of red/green fluorescence (Fig. 3Aii). Quantitative image analysis revealed oxidative stress-induced ΔΨm loss, with the red/green fluorescence ratio decreasing by 4.72 ± 0.47 (2.0 mM H₂O₂) and 0.94 ± 0.08 (3.0 mM H₂O₂). In 2.0 mM H₂O₂-treated cells, mitochondria transplantation significantly restored the red/green fluorescence ratio to 6.15 ± 0.338 (p = 0.002 vs. stressed controls; p = 0.486 vs. untreated controls), demonstrating complete recovery to baseline levels with no significant difference from normal conditions. TM cells subjected to severe oxidative stress (3.0 mM H₂O₂) showed irreversible ΔΨm impairment, with mitochondria supplementation achieving only partial recovery (Red/Green fluorescence ratio: 1.88 ± 0.21 vs. 6.15 ± 0.34 in untreated controls; p < 0.001). This therapeutic limitation contrasted sharply with the robust recovery observed under moderate stress (2.0 mM H₂O₂), establishing a clear efficacy threshold for mitochondria transplantation. Flow cytometric quantification of JC−1 fluorescence ratios (Fig. 3Aiii) corroborated the above findings, demonstrating a dose-dependent shift from red to green fluorescence under H₂O₂-induced oxidative stress. Following exogenous mitochondria transplantation, a significant increase in the proportion of cells with high red fluorescence was observed, indicating an elevated population of cells with higher mitochondrial membrane potential, from 35.8%±1.5% to 82.8 ± 1.2% in 2.0 mM H₂O₂-treated cells. The differential outcomes correlated with our uptake experiments that moderate stress enhanced mitochondria incorporation, while excessive damage likely overwhelmed cellular quality control mechanisms needed for functional integration.

Fig. 3.

Assessment of TM cells function following mitochondrial transplantation. (Ai-Aiii) Mitochondrial membrane potential (ΔΨm) in TM cells via JC-1 assay. (Ai) CLSM images of JC-1-stained TM cells. CCCP, a kit-provided positive control, collapses ΔΨm. Red fluorescence indicates JC-1 aggregates, means high ΔΨm, while green represents JC-1 monomers, means low ΔΨm (Scale bar = 10 μm). (Aii) Quantification of JC-1 fluorescence by flow cytometry. Red dots: JC-1 aggregates; Blue dots: JC-1 monomers. (Aiii) Quantitative analysis of JC-1 aggregate/monomer ratio according to the flow cytometry results (n = 3). (Bi-Biii) ‌ Intracellular ROS levels in TM cells.(Bi) CLSM images of ROS in TM cells. Blue, nuclei; green, ROS (Scale bar = 20 μm). (Bii) Flow cytometric quantification of ROS. (Biii) Mean fluorescence intensity of ROS (n = 3). (Ci-Ciii) Mitochondrial superoxide levels in TM cells. (Ci) CLSM images of mitochondrial superoxide in TM cells. Blue, nuclei; red, superoxide, green, mitochondrial (Scale bar = 10 μm). (Cii) Flow cytometric quantification of mitochondrial superoxide in TM cells. (Ciii) Mean fluorescence intensity of superoxide (n = 3). Effect of mitochondrial transplantation on (D) ATP content, (E) NAD+/NADH and on (F) GSH/GSSG in TM cells (n = 3). (Gi) Annexin V/PI assay of apoptosis in TM cells after mitochondrial supplementation treatment using flow cytometry. Flow cytometry analysis was performed using standard compensation settings to ensure the accuracy of cell population analysis. (Gii) Quantitative analysis of cell survival rate based on the results of Gi (n = 3). Data are expressed as mean ± standard deviation. Statistical significance was calculated by ordinary one-way ANOVA

Intracellular ROS levels were quantified using the fluorescent probe DCFH-DA, which undergoes oxidation to fluorescent DCF in the presence of reactive oxygen species. Untreated control cells exhibited minimal cytoplasmic green fluorescence (Fig. 3Bi), reflecting low baseline oxidative stress. H₂O₂ treatment induced dose-dependent ROS accumulation, demonstrating ROS accumulation proportional to oxidant stress. Mitochondria supplementation effectively normalized ROS levels in cells exposed to 1.0–2.0 mM H₂O₂, reducing fluorescence to near-baseline values. However, this therapeutic effect was attenuated under severe oxidative stress (3.0 mM H₂O₂), where the fluorescence intensity remaining persistently elevated and showing no significant difference from pre-treatment levels. Flow cytometric analysis confirmed these observations (Fig. 3Bii-Biii), in the 2.0 mM H₂O₂ group, mitochondria transplantation significantly reduced the mean intracellular ROS fluorescence intensity from 73 ± 3.61 to 6.97 ± 0.21 (p < 0.001), a level comparable to that of the control group, demonstrating that exogenous mitochondria restored redox homeostasis in moderately stressed cells (2.0 mM H₂O₂) but showed limited efficacy against extreme oxidative damage (3.0 mM H₂O₂). These findings highlighted that both the therapeutic potential and functional limits of mitochondria transplantation in oxidative stress mitigation.

To specifically quantify mitochondrial superoxide (O₂•⁻) production, we employed a mitochondria-targeted fluorescent probe that selectively detects intra-mitochondrial O₂•⁻ without cross-reactivity to other ROS species. After labeling endogenous mitochondria in TM cells with a green fluorescent marker, oxidative stress was induced followed by therapeutic interventions. Untreated control cells exhibited minimal baseline red fluorescence (Fig. 3Ci), whereas H₂O₂ exposure induced dose-dependent superoxide (O₂•⁻) accumulation, reaching peak levels at 3.0 mM H₂O₂ concentration. Mitochondria transplantation effectively restored normal O₂•⁻ levels in TM cells exposed to 1.0–2.0 mM H₂O₂, with only residual fluorescence detectable. However, this therapeutic effect was significantly attenuated under severe oxidative stress (3.0 mM H₂O₂), where O₂•⁻ levels remained markedly elevated. Flow cytometric analysis confirmed these findings (Fig. 3Cii-iii), showing parallel trends to our total ROS measurements. These results demonstrate that exogenous mitochondria can scavenge mitochondrial O₂•⁻ and restore redox homeostasis, but their efficacy is constrained under extreme oxidative damage.

Mitochondria, as the central organelles of cellular energy metabolism, play a critical role in ATP production through oxidative phosphorylation. The measurement of ATP levels serves as a key indicator for evaluating mitochondrial functional status. Quantification of intracellular ATP content revealed significant metabolic impairment following oxidative stress (Fig. 3D). H₂O₂ exposure reduced ATP levels from 7.369 ± 0.0056 µM/mg (untreated controls) to 4.758 ± 0.0041 µM/mg in the 3.0 mM treatment group (p < 0.01). Mitochondria transplantation restored ATP production in a stress-dependent manner, with maximal recovery observed in the 2.0 mM H₂O₂ group (7.587 ± 0.0009 µM/mg, vs. untreated controls, p < 0.001), demonstrating near-complete metabolic rescue under moderate oxidative stress.

The GSH/GSSG and NAD+/NADH ratios serve as key indicators of cellular redox homeostasis, where elevated ratios reflect reduced oxidative stress and proper metabolic function, while decreased ratios indicate oxidative damage and impaired energy metabolism, together providing critical insights into cellular health and stress responses [36, 37]. The GSH/GSSG system primarily serves as the cellular ‘antioxidant and detoxification system’; restoration of this ratio signifies the recovery of the cell’s capacity to scavenge reactive oxygen species (ROS) and maintain a reduced intracellular environment. In contrast, the NAD+/NADH system functions as the central ‘energy metabolism and signaling hub’; its normalization indicates the restoration of core bioenergetic pathways, including mitochondrial respiratory chain function and glycolytic/TCA cycle flux. As shown in Fig. 3E-F, in the control cells, the GSH/GSSG and NAD⁺/NADH ratios were 3.54 ± 0.05 and 38.77 ± 1.46, respectively. Treatment with 2.0 mM H₂O₂ significantly decreased these ratios to 1.39 ± 0.21 and 2.58 ± 0.11, indicating severe oxidative disruption. Mitochondria transplantation markedly increased the GSH/GSSG ratio to 34.04 ± 0.5, demonstrating a substantial recovery of the intracellular redox buffer capacity (p = 0.002 vs. control group). Furthermore, the NAD⁺/NADH ratio was restored to 35.8 ± 1.1, showing no significant difference compared to the healthy control group (p = 0.619). This indicated that mitochondrial respiratory chain function and central metabolic pathways were effectively reinstated. The results in the mild treatment group (1.0 mM H₂O₂) were consistent with the overall findings. However, this therapeutic effect was attenuated under severe oxidative damage (3.0 mM: GSH/GSSG = 0.75 ± 0.07, NAD+/NADH = 1.87 ± 0.12), establishing a clear efficacy threshold that aligned with our bioenergetic and ROS measurements.

Finally, we evaluated the impact of exogenous mitochondria transplantation on cell viability. Similarly, we first induced oxidative stress damage at low, moderate and high gradients using H₂O₂ respectively, then treated them using exogenous mitochondrial supplementation, and finally assessed cell survival rates via flow cytometry. As shown in Fig. 3 Gi-Gii, it could be seen that the apoptosis rate induced by H₂O₂ was concentration-dependent. For the low and middle groups, the cell survival rate could be significantly improved after mitochondria treatment, especially in the 2.0 mM H₂O₂ group, the cell survival rate was significantly increased from 80.93%±0.54%to 91.57 ± 0.29% (p < 0.001). However, the excessive cell damage in the 3.0 mM H₂O₂ group may exceed the mitochondrial repair capacity, leading to irreversible collapse of metabolic pathways and the reason for lower cell survival.

These findings demonstrate that exogenous mitochondria can functionally restore redox balance and bioenergetics as well as the cell viablity following moderate oxidative injury (optimal at 2.0 mM H₂O₂).

Mitochondrial transplantation improves viability through mitophagy suppression

Thus far, we have fully demonstrated that mitochondria supplementation therapy can effectively restore cellular function in oxidative stress-injured TM cells. To further validate whether this remarkable therapeutic potential is attributable to the supplemented exogenous mitochondria, we subsequently performed proteomics and WB experiments for verification.

TM cells were treated with 2.0 mM H₂O₂ followed by administration of two concentrations of mitochondria supplementation (2 and 4 µg protein/ mL), and then subjected to proteomic analysis. Hierarchical clustering revealed significant alterations in protein levels following H₂O₂-induced damage and subsequent mitochondria therapy (Fig. 4A). Compared to the control group, 2.0 mM H₂O₂ treatment induced 1,033 differentially expressed proteins (DEPs), including 437 up-regulated and 596 down-regulated proteins. Relative to the 2.0 mM H₂O₂-treated group, mito-1 treatment resulted in 541 DEPs (540 up-regulated and 1 down-regulated proteins), while mito-2 treatment yielded 531 DEPs (530 up-regulated and 1 down-regulated proteins). Gene ontology (GO) biological process enrichment analysis of the DEPs between 2.0 mM H₂O₂ and 2.0 mM H₂O₂ +mito groups indicated that the up-regulated proteins were predominantly localized to mitochondria and primarily associated with energy generation-related processes, such as “ATP generation process”, “mitochondrial complex assembly”, “oxidative phosphorylation” and “fatty acid metabolic process” (Fig. 4B).

Fig. 4.

(A) Hierarchical clustering of differentially expressed proteins in control, 2.0mM H₂O₂ TM cells and treated with 2–4 µg mitochondria. (B) Gene Ontology (GO) analysis of differentially expressed proteins in 2.0mM H₂O₂ TM cells vs. treated with 2 µg mitochondria. (C) Hierarchical clustering of differentially expressed mitochondria-related proteins in control, 2.0mM H₂O₂ TM cells and treated with 2–4 µg mitochondria. (D) Gene Ontology (GO) analysis of differentially expressed mitochondria-related proteins in 2.0mM H₂O₂ TM cells vs. treated with 2 µg mitochondria. (Ei) Western blot (WB) analysis was performed to detect apoptosis-related proteins, including BAX, BCL-2 and cleaved caspased-3. (Eii-Eiv) Relative protein expression levels of (Eii) cleaved caspased-3, (Eiii) BAX and (Eiv) BCL-2. (Fi) Western blot (WB) analysis was performed to detect mitophagy-related proteins, including P62, Parkin, PINK-1 and LC3B. (Fii-Fv) Relative protein expression levels of (Fii) P62, (Fiii) PINK-1, (Fiv) Parkin and (Fv) LC3B-II/I. (Gi) Immunofluorescence staining of LC3 and TOM20 in TM cells. Scale bar, 10 μm. And (Gii) the number of LC3 puncta in cells. (H) Illustration showing the possible therapeutic mechanism pathway by which mitochondrial supplementation restores mitophagy levels to normal following H₂O₂-induced damage in TM cells, leading to balanced oxidative stress and ultimately promoting cell viability. Data are expressed as mean ± standard deviation. Statistical significance was calculated by ordinary one-way ANOVA

Subsequently, we isolated mitochondria-associated DEPs and generated interaction heatmaps coupled with GO biological process enrichment analysis. As illustrated in Fig. 4C-D, mitochondria supplementation therapy significantly up-regulated functional mitochondrial proteins, enhancing ATP production, antioxidant capacity, and ROS homeostasis. Notably, we observed overexpression of the mitophagy-related protein TOMM20 and increased levels of anti-apoptotic proteins such as BCL-2.

To validate these findings, we performed western blotting to quantify key apoptotic and mitophagy-related proteins (Fig. 4E-F). Compared to the control group, 2.0 mM H₂O₂ treatment markedly elevated apoptotic proteins (BAX and cleaved caspase-3) while down-regulating anti-apoptotic BCL-2. Mitochondrial intervention reversed these effects, restoring BCL-2 to normal levels and significantly suppressing BAX and cleaved caspase-3 expression (BCL-2: 0.171 ± 0.017 vs. 2.0mM, p < 0.001; BAX: 0.5221 ± 0.007 vs. 2.0mM, p < 0.001; cleaved caspase-3: 0.957 ± 0.07 vs. 2.0mM, p < 0.001). Additionally, we assessed mitophagy levels by detecting key mitophagy-related proteins in the canonical Parkin/PINK1 pathway. Compared with the control group, 2.0 mM H₂O₂-induced oxidative damage triggered aberrant mitophagy activation, characterized by increased expression of P62、Parkin, PINK1 and LC3B. Following mitochondrial supplementation, mitophagy was restored to homeostasis, maintaining levels comparable to baseline conditions. Co-localization of LC3 puncta with mitochondria (labeled by translocase of outer mitochondrial membrane 20, Tom20) serves as a marker for mitophagy [38]. In control group, the number of LC3 puncta co-localized with TOM20 was relatively low. Treatment with 2.0 mM H₂O₂ significantly increased the number of co-localized puncta, while subsequent mitochondrial supplementation therapy reduced the number of co-localized puncta to a level that showed no statistically significant difference from the control group. Therefore, based on the results, we concluded that mitochondria supplementation therapy restores cellular homeostasis by normalizing the aberrantly elevated mitophagy, thereby improving mitochondrial function, rebalancing oxidative stress, enhancing ATP production and ultimately promoting cell proliferation (Fig. 4G).

Assessment of exogenous mitochondrial distribution in TM 3D models and therapeutic effects in vitro

To better mimic the three-dimensional structure of trabecular meshwork (TM) tissue in vivo, we constructed 3D TM cell spheroids by in vitro culture. As illustrated in the schematic diagram of Fig. 5Ai-Aii, we adopted the hanging drop method to prepare spherical 3D TM cell spheroids. Subsequently, the exogenous mitochondria were labeled with MitoTracker Deep Red 633, and co-culture with 3D spheroids. Finally, the spatial distribution of exogenous mitochondria within 3D spheroids could be visualized via ‌confocal microscopy with 3D tomography.‌ As shown in Fig. 5B, faint red fluorescence was detectable deep within the spheroids, indicating that exogenous mitochondria could penetrate and be internalized by TM cells in the 3D models, possibly relying on concentration gradient diffusion. Figure 5C-E illustrated the distribution of exogenous mitochondria in spheroids subjected to varying degrees of oxidative damage. It could be clearly seen that spheroids exposed to low and moderate H₂O₂ concentrations (1.0–2.0 mM) exhibited significantly stronger red fluorescence intensity and penetration depth compared to the high-damage group (3.0 mM). The 2.0 mM H₂O₂ group displayed the most robust fluorescence, consistent with observations in 2D monolayer cultures (Fig. 2Hi-Hii, Supplementary movie1-4), suggesting that moderate oxidative damage (2.0 mM H₂O₂) enhanced exogenous mitochondrial uptake and permeation in both 2D and 3D TM models, thereby maximizing therapeutic efficacy. Subsequently, we evaluated the therapeutic efficacy of mitochondrial treatment in ameliorating oxidative redox damage within the 3D spheroids. Initial assessment of ATP levels revealed a significant decrease in cellular ATP content following H₂O₂ treatment (Fig. 5F). The 2.0 mM H₂O₂ was identified as inducing moderate cellular damage. When treated with mitochondria, the ATP levels of each group increased, with the 2.0 mM treatment group exhibiting complete recovery to baseline ATP concentrations. Further analysis of redox homeostasis markers (GSH/GSSG and NAD⁺/NADH) through Fig. 5G-H showed that mitochondrial therapy effectively restored redox balance in low- and medium-concentration damage groups (1.0–2.0 mM H₂O₂), achieving near-normal redox status. However, the therapeutic efficacy was diminished in 3.0 mM H₂O₂ group, indicating concentration-dependent limitations in recovery capacity. Finally, cell viability assessments (Fig. 5I) demonstrated remarkable therapeutic outcomes in the 2.0 mM group, with survival rates increasing from 71.64 ± 1.31% to 98.67 ± 0.7% following mitochondrial treatment (vs. control, p=0.3). This concentration cohort showed optimal therapeutic response, suggesting 2.0 mM H₂O₂-induced damage represents the most effective treatment window for mitochondrial intervention.

Fig. 5.

Evaluation of mitochondrial supplementation therapy in 3D cell spheroids. (Ai) Schematic of TM cell spheroids prepared by the droplet method. (Aii) Schematic diagram of Z-axis imaging of spheroids via CLSM. (Aiii) Schematic of cell spheroids dispersed into a single-cell suspension. (B–E) CLSM images showing exogenous mitochondria uptake by spheroids under different conditions (mitochondrial protein concentration: 2 µg proteins/mL). (B) Uptake by untreated TM spheroids. (C) Uptake by spheroids pre-treated with 1.0 mM H₂O₂. (D) Uptake by TM spheroids pre-treated with 2.0 mM H₂O₂. (E) Uptake by TM spheroids pre-treated with 3.0 mM H₂O₂. Blue, nuclei; red, mitochondrial; green, TM cells (Scale bar = 100 μm). Effect of mitochondrial transplantation on (F) ATP content, (G) NAD+/NADH, (H) GSH/GSSG and on (I) survival rate in TM 3D spheroids (n = 3). Data are expressed as mean ± standard deviation. Statistical significance was calculated by ordinary one-way ANOVA

These results demonstrated that exogenous mitochondria effectively penetrated the interior of 3D cellular spheroids and the mitochondria supplementation therapy enhanced 3D spheroids viability by mitigating oxidative stress levels.

Exogenous mitochondria specifically enriched in TM tissues in vivo OHT models

To investigate the intraocular distribution of mitochondria following anterior chamber injection, we first labeled mitochondria with Mito-Tracker Green (1 µM, 37 °C, 10 min). As shown in schematic diagram 6 A, labeled mitochondria were slowly injected into the anterior chamber of both eyes of mice using a 33G needle with a Hamilton glass microsyringe a 45°angle directed toward the corneal limbus. After 24 h, eyes were harvested and cryosectioned to visualize exogenous mitochondrial distribution. ‌As shown in Fig. 6B, ‌mitochondrial intra-anterior chamber injection results in diffuse distribution across multiple ocular tissues. Distinct green fluorescence signals were observed in the cornea, ciliary body, trabecular meshwork, partial sclera and retina. Notably, intense fluorescence was detected in both the peripheral retina and inner retinal layers. These results preliminarily indicate that exogenously administered mitochondria diffuse into ocular tissues, likely mediated by concentration-dependent diffusion, primarily driven by passive diffusion mechanisms. Additionally, exogenous mitochondria exhibited pronounced enrichment in vascular-rich tissues.

Fig. 6.

Intraocular localization of exogenous mitochondria.‌(A)‌ Schematic of anterior chamber injection. Pre-labeled green fluorescent exogenous mitochondria were slowly infused into the anterior chamber via microsyringe at a 45℃near the limbus. ‌(B)‌ Distribution of exogenous mitochondria throughout the ocular globe. Blue, nuclei; green, mitochondrial. Red arrow: TM region. Yellow: Corneal region. White: Retinal region. (Ci) Enrichment of exogenous mitochondria in the TM region at different time periods. Blue, nuclei; green, mitochondrial (Scale bar = 100 μm). (Cii) Percentage of exogenous mitochondria in the TM region shown in Figure Ci (n = 3). ‌(Di)‌ Immunofluorescence analysis comparing endogenous vs. exogenous mitochondrial distribution in experimental groups. Blue, nuclei; red, endogenous mitochondrial, green, mitochondrial. (Dii) Percentage of exogenous mitochondria in the TM region shown in Figure Di (n = 3). TM, trabecular meshwork; SC, Schlemm’s canal; CB, ciliary body. Data are expressed as mean ± standard deviation. Statistical significance was calculated by ordinary one-way ANOVA

Subsequently, we examined the retention of exogenous mitochondria in the TM tissues after injection. As shown in Fig. 6Ci, 8 h post-injection, the exogenous mitochondria were detectable in the ciliary body, sclera and TM tissues. With the extension of time, the exogenous mitochondrial distribution shifted significantly that the exogenous mitochondria became predominantly enriched in the TM tissues, with markedly reduced signals in the ciliary body and sclera. This TM-specific enrichment intensified further at 24 h, as evidenced by increased fluorescence intensity. The proportion of exogenous mitochondria in the TM gradually increased over time, reaching 25.83 ± 0.012% at 24 h and then decreased to 18.87 ± 0.004% at 48 h. 28 days post-transplantation, faint but detectable fluorescence signals from exogenous mitochondria were still observed in the TM region, demonstrating the prolonged retention capacity of transplanted mitochondria within the target tissue (Fig. 6Cii).

To further observe the difference in the distribution of exogenous mitochondria in different eyes and assess the integration of exogenous mitochondria in the mitochondrial network, the endogenous mitochondria were labeled with red fluorescence via immunofluorescence. As shown in Fig. 6Di-Dii, compared to healthy eyes and saline-treated eyes, OHT eyes exhibited pronounced exogenous mitochondrial accumulation in the TM tissues, the percentage of their exogenous mitochondria in the TM reached 21.27 ± 0.019%, which was 5.18 and 8.58 times higher than that of the healthy and saline groups, respectively, showing a clear feature of specific distribution. ‌Simultaneously, we found that the exogenous mitochondria would be progressively integrated into the mitochondrial network, exhibiting the highest total mitochondrial density in the TM regions within the DEX-treated groups.

These findings demonstrated that in OHT models, the exogenous mitochondria would be specifically targeted accumulation within the TM tissues and further integrated into the mitochondrial network. This dynamic process culminates in a significant augmentation of TM tissues’ mitochondrial content, establishing a critical structural foundation for restoring physiological function in this glaucoma-relevant tissue.

Mitochondrial transplantation restores mitochondrial homeostasis in TM and exerts efficient IOP-lowering in vivo OHT models

Our primary focus was determining whether transplanted mitochondria could achieve effective and sustained IOP reduction. In vivo experiments, we first established an OHT model, where DEX administration was discontinued upon IOP stabilization at 15.4 ± 0.2 mmHg while the exogenous mitochondria injected into the anterior chamber of the eyes (Fig. 7Aii). The IOP was measured daily during the first week and then every 2–3 days thereafter. From the IOP change curve (Fig. 7B), we could see that in either the healthy groups or saline-treated groups, IOP showed no significant fluctuations following exogenous mitochondria. In contrast, the DEX group exhibited a potent IOP-lowering effect that IOP decreased from 15 mmHg to 11.2 mmHg within 24 h post-injection of exogenous mitochondria. Surprisingly, these effect persisted stably for 25 days, with an average IOP of 10.6 mmHg. At the same time, we found that there was no significant change in IOP after the PBS injection treatment both in healthy group, Saline or DEX group, confirmed that anterior chamber injection’s neutrality. When transplanted with dMito, no IOP-lowering effect was observed, showing a statistically significant difference compared to the DEX+Mito group (p < 0.001) but no significant difference compared to the DEX + PBS group (p=0.98). With daily continuous administration, the mean IOP was maintained at 12.1 ± 0.3 mmHg in the ‌Timolol maleate eye drops group and 11.1 ± 0.3 mmHg in the Latanoprost eye drops group, stably higher than the DEX+Mito groups by 1.5 mmHg and 0.5 mmHg, respectively.

Fig. 7.

In vivo efficacy evaluation of mitochondria transplantation. (Ai) Schematic diagram of anterior chamber injection of exogenous mitochondria in mice (20 µg proteins/mL, 3µL per eye). (Aii) Schematic diagram of mouse OHT model establishment and treatment protocol. The OHT model was constructed using DEX both eyes. Treatment commenced on day 23 post-successful modeling, with serial IOP monitoring. Eyes were removed on day 51 for HE and other efficacy assessments. (B) IOP curve, n = 10, 20 eyes. (C) Cumulative IOP values per eye over 51 days, n = 10, 20 eyes. (D) IOP difference between day 0 and day 51, n = 10, 20 eyes. (E) Body weight change curve in mice, n = 10. (Fi) TEM images of TM tissue from different treatment groups on day 51. Green arrows: healthy mitochondria. Yellow arrows: damaged mitochondria. (Fii) Quantification of healthy mitochondria in Figure Fi, n = 3. (Gi) TEM images of TM tissue in the DEX+Mito group at different time points (Scale bar = 500 nm). Green arrows: healthy mitochondria. Yellow arrows: damaged mitochondria. (Gii) Quantification of healthy mitochondria in Figure Gi, n = 3. (Hi) Schematic diagram of TM tissue dissection from mouse eyes. Red arrow: TM region. (Hii) SEM images of TM tissue from healthy eyes and the DEX+Mito group at different time points (Scale bar = 50 μm). Blue arrows: normal TM bands. Red arrows: abnormal TM bands. dMito, disrupted mitochondria

Subsequently, the cumulative IOP in each groups were quantified (Fig. 7C), revealing that the DEX+Mito group exhibited significantly lower cumulative IOP compared to the DEX group, DEX + PBS group, or DEX + IOP-lowering eye drops groups (p < 0.001).There was no statistically significant difference in the cumulative IOP between the healthy groups and the saline-treated groups. Meanwhile, we quantified the IOP changes before and after one week of treatment, and as shown in Fig. 7D, only the DEX+Mito group had the largest IOP changes, about 5 ± 0.5 mmHg(vs. DEX group, p < 0.001). Safety evaluations revealed transient weight loss during OHT induction that normalized post-DEX cessation across all groups, with mitochondrial-treated animals displaying weight recovery trajectories paralleling healthy controls and no evidence of systemic toxicity (Fig. 7E).

To investigate whether exogenous mitochondria transplantation lowers IOP by improving mitochondrial integrity in the TM, we examined mitochondrial quantity and morphology in TM tissues using TEM. As shown in Fig. 7Fi, mitochondria in the healthy groups exhibited intact morphology with distinct double-membrane structures and well-defined cristae formed by inner membrane folding, while the saline groups showed no significant differences from healthy controls. In contrast, OHT model displayed marked mitochondrial edema in TM cells, characterized by organelle swelling and near-complete cristae dissolution. Following exogenous mitochondria supplementation, we observed significantly increased mitochondrial density in TM, with restored normal morphology and cristae structure, along with an elevated proportion of healthy mitochondria. Notably, two IOP-lowering eye drops groups demonstrated no significant differences in mitochondrial morphology or healthy mitochondrial count compared to the DEX group (DEX+Timool vs. DEX, p = 1.0; DEX+Latanoprost vs. DEX, p = 0.50) (Fig. 7Fii), suggesting their hypotensive mechanisms are independent of TM mitochondrial modulation. When transplanted with dMito, the level of healthy mitochondria in the TM tissues did not increase, remaining comparable to the DEX group (p = 0.52). These findings indicated that exogenous mitochondrial supplementation effectively restores TM mitochondrial populations in OHT models, potentially enabling normal aqueous humor outflow function. In addition, we observed differences in TM mitochondria at various post-treatment time points. As illustrated in Fig. 7Gi‌, compared to the DEX group, the DEX+Mito group exhibited a ‌modest increase‌ in healthy mitochondrial count within the TM at ‌24 hours post-intervention‌. Notably, these numbers ‌progressively increased over time‌, culminating in an ‌approximately threefold elevation‌ in healthy mitochondria by the ‌treatment endpoint (day 28)‌ relative to pretreatment levels (Fig. 7Gii).

Furthermore, following the isolation protocol illustrated in Fig. 7Hi, we dissected TM and employed SEM to evaluate microstructural alterations in TM. As shown in Fig. 7Hii‌, the healthy TM exhibited a distinct porous architecture with regularly arranged and smooth-surfaced trabecular bands. In contrast, DEX treatment induced significant structural disorganization characterized by obscured TM boundaries, roughened and irregularly aligned bands, and partial collapse of the sieve-like structures. Following mitochondria supplementation, progressive restoration of the TM microstructure was observed over time, manifested by gradually more distinct meshwork structure, smoother trabecular bands surfaces, and reestablished porous architecture. Based on the combined TEM and SEM results, we proposed that exogenous mitochondria supplementation may further activate endogenous cellular activity in TM cells, thereby promoting progressive and long-term structural restoration.

These findings demonstrated that in the OHT model, mitochondrial supplementation enables rapid, stable and sustained IOP lowing.‌ These therapeutic effect were mediated through increased healthy mitochondrial populations within the TM, which drives structural remodeling of the outflow pathway, culminating in normalized aqueous humor dynamics.

At the macroscopic level, we further employed HE staining to assess alterations in the anatomical structure of the TM. As illustrated in Fig. 8Ai, TM cells in the healthy and saline-treated groups exhibited orderly arrangements, with flattened, elongated nuclei stained deep purple. In the DEX-treated groups, including PBS and IOP-lowering drug treatment groups, the TM tissues displayed abnormally sparse intercellular spaces, disordered cell arrangements, and a relative reduction in cellular density within its region. As shown in Fig. 8Aii, following mitochondrial supplementation, cellular alignment in the TM area returned to normal, with a significant increase in cell numbers showing no discernible difference from the healthy groups (DEX+Mito vs. control group, p = 0.12). These results provided comprehensive evidence at both microscopic and macroscopic levels that mitochondrial supplementation therapy could reverse the structural alterations in the TM.

Fig. 8.

Effects of mitochondrial supplementation on TM function. (Ai) HE staining. The TM is outlined by the red dashed circle. (Aii) Quantification of TM area in Figure Ai HE images, n = 3. (Bi) Fluorescein sodium staining images of the ocular surface at different time points. (Bii) Aqueous humor outflow curve plotted based on fluorescence intensity in Figure Bi, n = 3

Finally, we conducted functional assessments by evaluating the aqueous humor outflow facility using fluorometric methods1. As shown in Fig. 8Bi-ii, the aqueous humor outflow rate in the DEX-treated group was significantly lower than that in the healthy group within 10 min, indicating impaired aqueous humor outflow function. However, after mitochondrial supplementation, the outflow rate approached normal levels, indicating that mitochondrial therapy significantly restored aqueous outflow function. This functional restoration primarily depended on the reversal of trabecular meshwork structural changes.

The IOP-lowering mechanism following mitochondria supplementation

How does mitochondria supplementation therapy induce structural alterations in the TM? To investigate this, we first collected aqueous humor from different treatment groups and we quantified ATP levels (Fig. 9A). As shown in Fig. 9B, DEX treatment significantly reduced ATP levels (4.94 ± 0.07) compared to healthy controls (7.38 ± 0.033; p < 0.001). Following mitochondria supplementation, ATP levels progressively increased with the duration of treatment. By 24 h post-treatment, ATP concentration rose to 5.56 ± 0.12 and continued to increase until reaching 7.11 ± 0.21 at the 80 h, showing no significant difference compared to the ATP level in the normal group (7.38 ± 0.033 vs. control, p = 0.018). Furthermore, the elevated ATP state persisted for up to 28 days.

Fig. 9.

(A) Schematic diagram of aqueous humor collection from mice; (B) ATP content in aqueous humor. (C) ROS levels and (D) 8-OHdG in aqueous humor measured by ELISA (n = 3). Confocal microscopy images of (E) PINK1and (F) Parkin immunofluorescence in TM tissue (Scale bar = 50 μm). Blue, nuclei; red, PINK; green, Parkin. Yellow arrow: TM region

Moreover, we quantitatively assessed oxidative stress levels in aqueous humor using ELISA to measure ROS and 8-OHdG concentrations. As demonstrated in Fig. 9C-D, the saline-treated and the control group exhibited low baseline oxidative markers, while both parameters were significantly elevated in the DEX-induced glaucoma model group (p < 0.01). Mitochondria transplantation therapy effectively reduced these oxidative indicators to near-normal levels, whereas PBS anterior chamber injection showed no therapeutic effect on either parameter.

Subsequently, we performed in vivo immunofluorescence to observe mitophagy in TM region. Using the classic mitophagy marker PINK1 and Parkin, we conducted semi-quantitative analysis of autophagic expression levels before and after treatment. Figure 9E-F demonstrated that healthy eyes exhibited low mitophagy levels in the TM, with only faint red fluorescence. Following DEX treatment, intense red fluorescence was observed in this region, indicating significantly enhanced mitophagy. After mitochondria supplementation, the red fluorescence intensity in the TM sharply decreased to levels comparable to the healthy group. These results demonstrated that mitochondrial therapy effectively reduced the ‌DEX-induced overexpression of mitophagy‌ in the TM, restoring it to normal levels. These finding were consistent with the ‌preceding Western blot observations‌.

Safety assessment

Our comprehensive safety evaluation demonstrated that mitochondria transplantation therapy exhibited excellent ocular and systemic biocompatibility (Fig. 10A-B). Ophthalmic examinations revealed no corneal edema and intact ocular surface integrity. Histopathological assessment of major organs showed preserved tissue architecture, with HE staining confirming normal morphology in heart (intact myocardial fibers), liver (no steatosis or inflammation), spleen (physiological red/white pulp distribution), lung (alveolar septa without thickening), and kidney (typical glomerular and tubular structures), further supporting the safety profile of this therapeutic approach.

Fig. 10.

Safety assessment. (A) Slit lamp photographs of the ocular surface. (B) H&E stained sections of liver, heart, spleen, lung and kidney (Scale bar = 100 μm)

Discussion

Mitochondrial damage is a critical driver of TM dysfunction in glaucoma, impairing oxidative phosphorylation, mitochondrial DNA integrity and mitophagy [13, 14]. Consequently, mitochondria-targeted therapies may represent a pivotal strategy for restoring TM homeostasis and halting disease progression. In this study, we aim to investigate whether mitochondria transplantation could restore TM function by replenishing mitochondrial mass and bioenergetics, thereby offering a novel therapeutic strategy for IOP control in glaucoma. Our findings demonstrate, for the first time, that exogenously transplanted mitochondria significantly enhance TM mitochondria function, revitalize cellular bioactivity, and promote structural remodeling, leading to sustained and stable IOP reduction. While prior studies have established the efficacy of mitochondria transplantation in restoring neuronal viability in central nervous system and myocardial ischemia disorders [39–41], this study provides the first experimental evidence demonstrating the efficacy of mitochondrial transplantation in a DEX-induced OHT model, specifically addressing TM dysfunction-a key pathological feature of glaucoma. Unlike neuronal studies focusing on acute neuroprotection, our results highlight the long-term structural and functional restoration of the TM, suggesting a unique role of mitochondria therapy in reversing glaucomatous damage rather than merely delaying degeneration. This divergence may stem from tissue-specific mitochondrial integration mechanisms or differences in the chronicity of disease progression between neurological and DEX-induced OHT model. Our study demonstrated that mitochondria transplantation exhibited significant therapeutic potential in DEX-induced OHT model characterized by mitochondrial dysfunction. It is important to clarify that this study was specifically conducted and validated in a DEX-induced OHT model, but does not yet establish neuroprotection or disease modification at the level of glaucomatous optic neuropathy. These findings not only expanded the application scope of mitochondria transplantation in ophthalmic diseases but also provided a theoretical foundation for its potential as a novel therapeutic strategy for other ocular conditions involving mitochondrial impairment.

How specific is the therapeutic effect of mitochondria transplantation? Our in vivo experiments revealed a critical finding: mitochondria transplantation exerted a robust IOP-lowering effect in the DEX-induced OHT model, yet showed no effect in healthy or saline-treated eyes. This indicated that the therapeutic action was not a nonspecific hypotensive effect but was strictly dependent on pre-existing TM damage, or more precisely, mitochondrial dysfunction. Under conditions without pre-existing mitochondrial dysfunction, the transplantation of healthy mitochondria elicited no significant positive or negative effects. Importantly, this outcome was specific to the administration of functional mitochondria; it was noted that transplantation of damaged or structurally compromised mitochondria could potentially trigger adverse immune or inflammatory responses [42]. These key observations underscored that mitochondrial transplantation acted as a targeted intervention to restore mitochondrial function in compromised TM cells, rather than functioning as a general hypotensive agent. While the current evidence was established in the DEX-induced OHT model, we concluded that future studies would be warranted to validate its applicability across other glaucoma paradigms.

Many studies have focused on pharmacological or antioxidant approaches to enhance TM mitochondrial function-primarily by activating endogenous repair mechanisms (e.g. PINK1/Parkin pathway)-these strategies often yield limited and transient benefits due to the inherently compromised mitochondrial function in diseased states [43, 44]. Notably, exogenous mitochondria transplantation not only directly increases mitochondrial numbers in target tissues but also functionally rescues and augments endogenous mitochondrial activity, enabling sustained production of healthy mitochondria. This represents a transformative ‘bio-augmentive’ strategy that fundamentally differs from conventional drug therapies by addressing both the symptom (IOP elevation) and root cause (mitochondrial failure) of glaucoma pathogenesis. Our study demonstrates that oxidative stress-damaged TM cells exhibit enhanced uptake of exogenous mitochondria, leading to increased intracellular mitochondrial density, and that were consistently observed in vivo in the DEX-induced OHT model where TM tissues with endogenous mitochondrial dysfunction showed significantly higher exogenous mitochondrial enrichment compared to controls (Figs. 2H-I, 5 and 6D). These observations are consistent with previous reports by other researchers, suggesting that damaged tissues may release specific factors to recruit exogenous mitochondria, preferentially localizing them to high-ROS regions [32, 45]. Alternatively, ATP depletion may activate the AMPK pathway, promoting macropinocytosis-mediated uptake of exogenous mitochondria, or the mitochondrial membrane potential gradient may drive this differential distribution [33, 46]. Importantly, while exogenous mitochondria distributed broadly across ocular tissues (sclera, ciliary body, retina), comprehensive safety assessments revealed no adverse effects, including absence of corneal edema, intraocular inflammation or systemic toxicity (Fig. 10). These findings collectively establish mitochondria transplantation as a precisely targeted and safe therapeutic strategy for oxidative stress-induced TM degeneration.

In the aqueous humor outflow pathway, the TM route accounts for approximately 60–90% of total aqueous drainage, playing a pivotal role in IOP regulation. In POAG, pathological alterations in TM structure lead to impaired aqueous outflow, resulting in aqueous accumulation in the anterior chamber and consequent IOP elevation [47, 48]. Therefore, therapeutic strategies targeting TM outflow function represent the most effective and specific approach for glaucoma treatment. In this study, we compared the therapeutic effects of mitochondria transplantation with two classical IOP-lowering drugs-timolol eye drops and latanoprost eye drops. Our results demonstrate that mitochondria transplantation exhibits superior advantages in terms of IOP reduction magnitude, onset time, duration of effect and dosing frequency, likely attributable to fundamental differences in the mechanisms of action. As a non-selective β-blocker, timolol primarily lowers IOP by suppressing aqueous humor production [49]. However, this mechanism shows limited efficacy in open-angle glaucoma (OAG) patients who typically maintain normal aqueous production capacity. The uveoscleral outflow pathway accounts for approximately 10–30% of total aqueous humor drainage. Latanoprost exerts its IOP-lowering effect primarily by enhancing this pathway, increasing its contribution to 35–50% in glaucoma [50]. While latanoprost treatment achieved a mean IOP of approximately 11.5 mmHg in our study, its effects were characterized by significant fluctuations and a relatively slow onset of action. Exogenous mitochondria transplantation demonstrated rapid, stable, and sustained IOP-lowering effects, reaching baseline levels by the second day post-injection and maintaining efficacy for nearly one month with just a single administration. Compared to the two aforementioned IOP-lowering medications, this approach showed significant advantages.

Within TM cells, mitochondria form interconnected networks regulated by finely tuned fusion-fission dynamics [51]. The fission process is governed by dynamin-related protein 1 (DRP1), a GTPase whose deficiency disrupts mitochondrial homeostasis [52–55]. Selective elimination of damaged mitochondria via PINK1/Parkin-dependent mitophagy is critical for maintaining TM cell viability. Emerging evidence suggests that impaired mitophagy in TM cells leads to pathological accumulation of reactive oxygen species (ROS) and nitrogen radicals, contributing to oxidative stress-mediated cell death [56, 57]. Our data reveal significant mitochondrial dysfunction in glaucomatous TM cells, characterized by elevated expression of mitophagy markers (PINK-1, Parkin, LC3B) alongside Caspase-3 activation, suggesting a nexus between bioenergetic failure and apoptotic signaling. Additionally, mitochondria transplantation markedly enhanced mitochondrial bioenergetics while concurrently suppressing both caspase-3-mediated apoptosis and Parkin-dependent mitophagy in TM cells of DEX-induced OHT mice. Following mitochondria supplementation, the PINK1-Parkin pathway exhibits differential regulation (either enhancement or suppression), primarily dependent on the treatment phase and cellular microenvironment. In the early phase (< 48 h post-transplantation), exogenous healthy mitochondria may activate the PINK1-Parkin pathway through two mechanisms: (1) triggering a “new-for-old” replacement effect to selectively clear residual damaged mitochondria (manifested as a transient increase in LC3B) and (2) indirectly enhancing endogenous mitophagic activity by restoring mitochondrial membrane potential (ΔΨm) and ATP supply. In the long-term phase (> 72 h), once healthy mitochondria fully replace damaged ones, the reduction in ROS and mtDNA leakage weakens PINK1-sensed damage signals, leading to decreased autophagic demand. Additionally, cell-type specificity plays a critical role: neurons with high basal autophagy levels are more likely to exhibit pathway activation, whereas fibroblasts with low basal mitophagy tend toward suppression [39, 56, 58]. This dynamic equilibrium reflects the cell’s precise regulation of mitochondrial quality-balancing the clearance of residual damage while avoiding energy waste from excessive mitophagy. In this study, mitochondrial uptake results (Fig. 3) demonstrate that intracellular healthy mitochondria are restored within 24 h, accompanied by normalized mitochondrial function (Fig. 4). Consequently, cellular demand for mitophagy decreases, maintaining lower mitophagic activity.

In the DEX model, the rapid mitochondrial damage may lead to mitophagy “overload,” consequently causing excessive enhancement of mitophagy levels. The observed “reduction” in mitophagy levels following mitochondria transplantation therapy may carry dual implications. This does not simply represent suppression of a beneficial quality control mechanism, but rather reflects an improvement in overall mitochondrial health. After supplementation with exogenous healthy mitochondria, the proportion of damaged mitochondria within cells significantly decreased, leading to a corresponding reduction in the “demand for mitophagy” required for their clearance. Therefore, we observed decreased mitophagy levels in TM after mitochondrial transplantation, which is a positive indicator of homeostatic recovery. Following transplantation of healthy mitochondria, we observed increased ATP levels, decreased oxidative stress, and enhanced cellular viability in TM cells (Fig. 3). This indicates an overall improvement in the functional mitochondrial pool within cells, with enhanced mitochondrial network function. The reduction in mitophagy levels is consistent with the observed therapeutic benefits.

While the current research on the mechanisms underlying mitochondrial transplantation is not yet fully elucidated, based on existing studies and our experimental results, we speculate that transplanted mitochondria may suppress excessive mitophagy through a dual synergistic mechanism, with the two aspects potentially complementing each other temporally and functionally. On the one hand, the direct integration of exogenous healthy mitochondria can rapidly replenish intracellular ATP and reducing equivalents (e.g., NADH/NADPH), thereby negatively regulating the overactivation of mitophagy. On the other hand, transplanted mitochondria may release certain molecules that act as signals, downregulating the overactivated PINK1/Parkin-mediated mitophagy pathway through intracellular pathways (AMPK) or intranuclear pathways (affecting the transcriptome) [59–61].

Current research on mitochondrial transplantation raises more questions than answers, primarily due to unresolved mechanisms and variable outcomes. A pivotal unanswered question is whether functional recovery stems from differences in mitochondrial internalization pathways or from dynamic interactions between exogenous and endogenous mitochondria within host cells-a process requiring further elucidation. Contradictory evidence exists regarding mitochondrial persistence: While Nicolás-Avila et al. reported macrophage-mediated degradation of cardiomyocyte-derived mitochondria within days [62], Lin et al. have detected that transplanted mitochondrial signals in recipient neurons for up to 4–12 weeks [58]. Our data reveal that the fluorescent signal of exogenous mitochondria persist in TM tissues for at least 4 weeks, suggesting that TM cells may uniquely evade mitochondrial degradation. Based on the observation by Doulamis et al. that exogenous mitochondria can survive and remain functional in cardiac muscle for up to 28 days [63], we believe the duration of sustained fluorescence observed in the eye in this study most likely represents the retention period of exogenous mitochondria within the eye. However, whether exogenous mitochondria functionally integrate with endogenous mitochondrial networks in TM cells remains unknown, warranting investigation into potential inter-organelle communication (e.g., via mitochondrial fusion or metabolite exchange).

The integration capacity of exogenously transplanted mitochondria remains controversial and appears to be cell-type-dependent. While studies supporting the non-integration hypothesis demonstrate rapid degradation of exogenous mitochondria via PINK1-Parkin-mediated mitophagy in highly phagocytic cells like macrophages, evidence for the integration hypothesis shows that neurons and cardiomyocytes-with high mitochondrial fusion activity-can maintain transplanted mitochondria long-term through upregulation of fusion proteins [62]. Our observation of 28-day mitochondrial persistence in trabecular meshwork (TM) cells suggests potential immune privilege (e.g., low CD36 expression), but definitive proof of functional integration requires further validation using live-cell imaging techniques.

As shown in Fig. 7Gi-ii, we observed that while the number of healthy mitochondria in TM cells remained relatively low on the first day after exogenous mitochondrial transplantation, it increased significantly over time, reaching approximately three times the pre-treatment level by day 28. This intriguing phenomenon, combined with our other findings, we proposed that the initial transplantation replenished the intracellular mitochondrial pool. The limited number of healthy mitochondria introduced appeared to activate or induce endogenous mitochondrial biogenesis, thereby restoring mitochondrial homeostasis and mitophagy, enhancing cellular bioenergetics, and ultimately stimulating the production of additional healthy mitochondria, which establishing a positive feedback loop. Regarding the specific underlying mechanism, we supported the view proposed by Liu et al. that mitochondrial genome transfer might promote mitochondrial biogenesis and energy metabolism in recipient cells [64].

Our study utilized mitochondria isolated directly from mouse liver tissue, which precluded the use of genetic engineering approaches to express encoded mitochondrial reporter proteins (e.g., mito-GFP) in donor cells. To minimize potential dye leakage, we carefully optimized the concentration and incubation time of MitoTracker dyes to avoid over-labeling and performed thorough washing steps after mitochondrial extraction to remove unbound free dye. Furthermore, we confirmed that the fluorescence intensity of the dye-labeled mitochondria remained stable within the cells for up to 48 h (Supplementary Fig. 4). However, we cannot entirely rule out the possibility of dye leakage beyond this timeframe, which represents a limitation of the current study.

Although this study demonstrated the promising therapeutic potential of mitochondria transplantation in a mouse model of DEX-induced OHT, its clinical translation faces several important challenges. First, the source of mitochondria represents a key issue: while autologous mitochondria could avoid immune rejection, obtaining a sufficient quantity of functional mitochondria from glaucoma patients (who are typically elderly) may prove difficult; whereas allogeneic mitochondria transplantation requires consideration of immune matching and long-term safety concerns. Second, as a chronic progressive disease, glaucoma may require repeated intraocular injections, which increases procedural risks and patient burden. Furthermore, as a short-term animal study, our experiment may not fully reveal long-term immune responses or chronic inflammation risks. Future research should further explore stable mitochondrial sources, optimize delivery methods, and conduct long-term safety assessments in larger animal models.

In summary, our study demonstrates that mitochondrial transplantation represents a transformative strategy for glaucoma therapy by directly addressing the root cause of trabecular meshwork (TM) dysfunction—mitochondrial failure. Unlike conventional IOP-lowering drugs that provide transient symptomatic relief, mitochondrial augmentation achieves sustained TM functional recovery through restoration of bioenergetics (enhanced ATP and lowered ROS) and dynamic regulation of mitophagy. These findings position mitochondrial therapy not merely as an alternative treatment but as a potential disease-modifying intervention for refractory glaucoma.

Methods

Ethics declarations

All animal studies performed in this study were approved by the Animal Experimental Ethics Committee of Wenzhou Medical University (Approval number: YSG25080102) and conducted in accordance with the committee’s guidelines.Six-week-old male C57BL/6J mice were provided by Zhejiang Provincial Animal Center (Zhejiang, China). Human trabecular meshwork (TM) cells were obtained from Wuhan Pricella Biotechnology Co., Ltd. (Wuhan, China).

Cell culture

Cells were cultured in a complete medium composed of Dulbecco’s Modified EagIe’s Medium: F-12 (DMEM/F12) (Cytiva, HyClone Laboratories, Logan, UT, USA) containing 10% foetal bovine serum (FBS) (Gibco, Life Technologies Corporation, NY, USA) and 1% penicillin-streptomycin solution (Thermo Fisher Scientific, Waltham, USA). Cells were expanded in the incubator with 5% CO2 at 37 °C and used for the experiments at least 1 week after culture at 37 °C.

Mitochondria isolation

Mitochondria isolation was performed according to the manufacturer’s protocol of mitochondrial isolation and purification kit (Beyotime Biotechnology, Haimen, China) (Fig. 2A). Briefly, healthy C57BL/6 mice weighing about 20–22 g were taken, collect liver tissue after euthanasia and wash with pre-cooled physiological saline to remove hair and blood residues. Under ice bath conditions, cut the liver tissue into 5 mm2 pieces and transfer to a homogenizer. Add mitochondrial isolation reagent at 1 mL per 200 mg of tissue, then homogenize on ice until the solution becomes smooth and particle-free. Centrifuge at 400×g for 5 min at 4 °C and repeat twice to remove unbroken cells. Collect the supernatant and centrifuge at 10,000×g for 10 min at 4 °C. Finally, resuspend the pellet in ice-cold PBS and centrifuge again at 12,000×g for 10 min at 4 °C to obtain purified mitochondria. The purified mitochondria were weighed, and the extraction rate was calculated using the following formula: Mitochondrial extraction rate (%) = Purified mitochondria (mg) / Liver tissue (mg)× 100%.

These obtained mitochondrial precipitations were resuspended in store buffer and used immediately.

To generate structurally disrupted mitochondria (dMito), the isolated mitochondria were subjected to ultrasonic treatment for 20 min.

Mitochondrial proteins were centrifuged at 12,000 g for 10 min after using the lysate. Collect the supernatant and quantify the protein concentration using the Bradford method.

Characterization of isolated mitochondria

Mitochondria were resuspended in PBS and the particle size distribution was performed on the Malvern Instrument (Zetasizer Nano ZS90,Malvern, UK). Fixed the mitochondria with 2.5% glutaraldehyde for 12 h at 4 °C and the morphologies of mitochondria was observed using transmission electron microscopy (TEM) (HT7800/HT7700, HITACHI, JP). The number of mitochondria with healthy morphology was counted, and the area and perimeter were quantified using ImageJ (three independent sections for each group).

Mitochondrial membrane potential was assessed using MitoTracker Deep Red 633 and JC-1 to evaluate mitochondrial activity and functionality. Briefly, for MitoTracker staining, mitochondrial suspensions (protein concentration: 0.5 mg/mL) were incubated with MitoTracker™ Deep Red 633 solution (100 mM) at 37 °C for 10 min, the excess dye solution was carefully removed and gently washed 3–4 times with PBS, and then visualized using confocal laser scanning microscopy (CLSM) (Zeiss Germany). For JC-1 detection, mitochondrial suspensions (protein concentration: 0.5 mg/mL) were mixed with JC-1 working solution according to the manufacturer’s instruction incubated at 37 °C for 10 min. The mitochondrial were then collected by centrifugation at 12,000 g for 10 min at 4 °C, resuspended in a minimal volume of PBS, and mounted on glass slides for CLSM analysis. Alternatively, the samples were transferred to a 96-well plate and loaded into a microplate reader for fluorescence spectrum scanning, with excitation and emission wavelengths set at 485 nm and 590 nm, respectively. For positive controls, mitochondrial samples were pretreated with CCCP solution (10 µM) for 10 min.

TM cells oxidative stress injury model

A classical H₂O₂-induced oxidative stress model was established. The procedure is as follows: TM cells were seeded in 96-well plates at a density of 5 × 10³cells per well and cultured overnight. Then the cells were treated with a series of H₂O₂ concentrations (0.2-5.0 mM) for 12 h. Followed by washed with three times PBS, the cells were further cultured for 12 h in fresh medium. 20µL MTT solution was added to each well. After 4 h of incubation, 160 µL DMSO was added to dissolve formazan crystals. Finally, the optical density (OD) was measured using a microplate reader (SpectraMax 190,USA). For the control group it was not treated with H₂O₂. The cell viability was calculated by OD (sample)/OD (control)×100%. Triplicate experiments were performed (n = 3).

Mitochondrial uptake

TM cells were seeded in 24-well plates at a density of 5 × 10⁴ cells/well and allowed to adhere overnight. Subsequently, TM cells were pre-labeled with CFDA-SE, and the exogenous mitochondria (mitochondrial protein concentration:2 µg proteins/mL) were pre-labeled with Mito-Tracker Deep Red 633, then gently washed to remove excess dye before being added to TM cells. At 1, 2, 12, and 24 h post-incubation, nuclei were stained with Hoechst. Mitochondrial uptake was visualized using CLSM.

To examine the differences in mitochondrial uptake by TM cells under different oxidative stimuli, we screened 1.0, 2.0, and 3.0 H₂O₂ as the low-medium-high damage conditions respectively, according to the MTT results. Briefly, TM cells were pre-treated with these H₂O₂ concentrations for 12 h. Following mitochondrial labeling (mitochondrial protein concentration: 2 µg proteins/mL), uptake was observed after 12 h via confocal microscopy.

For quantitative analysis, flow cytometry (BD Biosciences, CA, USA) was used to measure fluorescence intensity of labeled mitochondria internalized by TM cells under different oxidative conditions.

We used TEM method to observe the distribution and morphology of exogenous mitochondria in TM cells.The steps were as follows: TM cells were seeded in 6-well plates at 1 × 10⁶ cells/well and allowed to adhere overnight. Cells were pretreated with 1.0, 2.0, or 3.0 mM H₂O₂ for 12 h, washed with PBS, and incubated with exogenous mitochondria (mitochondrial protein concentration:) for 12 h. Finnally, the cells were trypsinized, collected after centrifugation and then fixed with 2.5% glutaraldehyde at 4 °C for 24 h. After dehydration and sectioning, mitochondrial distribution and morphology were observed using TEM.

Mitochondrial membrane potential detection

TM cells were seeded in 24-well plates at a density of 5 × 10⁴ cells/well. After 24 h of cell adhesion, the cells were exposed 2.0, and 3.0 H₂O₂ for 12 h. Then washed three times with PBS, and incubated with exogenous mitochondria (mitochondrial protein concentration:2 µg proteins/mL) for 12 h. Subsequently, incubate cells with JC-1 working solution at 37 °C for 30 min in the dark. Finally, the cells were fixed with 4% paraformaldehyde and stained with DAPI for nuclear visualization for CLSM observation. TM cells without any treatment or treated with CCCP were used as control and positive control groups, respectively. ImageJ was used to perform semi-quantitative measurements of red and green fluorescence intensities, followed by calculation of the red/green fluorescence ratio.

For quantitatively analysis, the cells were stained by JC-1 and collected for flow cytometry. The proportions of high red fluorescence and high green fluorescence were separately quantified.

ROS and mitochondrial ROS determination

Briefly, cells were seeded in 24-well plates at a density of 5 × 10⁴ cells per well and treated with different concentrations of H₂O₂ solution (1.0, 2.0 or 3.0 mM H₂O₂). For the mitochondrial treatment group mitochondria were added (mitochondrial protein concentration:2 µg proteins/mL). After 12 h of incubation, intracellular ROS was labeled with DCFH-DA, nuclei were stained with Hoechst, and ROS levels were visualized using CLSM. For quantitatively analysis, the cells were stained by DCFH-DA and collected for flow cytometry.

To further investigate changes of ROS in mitochondria, we selectively measured mitochondrial superoxide levels using a mitochondrial superoxide assay kit with MitoSO™Red. The procedure is as follows: TM cells were exposed to H₂O₂ followed by mitochondrial treatment. After 12 h incubation, cells were washed three times with PBS and then stained with Mito-Tracker Green (1 µM, 37 °C, 10 min) to label mitochondria and the excess dye solution was carefully removed and gently washed 3–4 times with PBS. Subsequently, 5 µM MitoSO™ Red solution was added at 37 °C for 20 min. The nuclei were stained with Hoechst stain and observed under CLSM.

ATP determination

Measurement of the intracellular or aqueous humor ATP levels using an ATP assay kit according to the manufacturer’s instructions. Typically, the cells are oxidatively damaged by H₂O₂. For the mitochondria treatment group mitochondria were added (mitochondrial protein concentration: 2 µg proteins/mL). After 12 h of incubation, cells were washed, lysed and centrifuged at 12,000 g for 10 min at 4 °C. The supernatants were collected for photochemical reaction with ATP working solutions. The ATP standard solutions were prepared according to the kit instructions. Measure the luminescence of samples or ATP standard solutions by a luminescence (SpectraMax M5, USA). The intracellular ATP levels were then calculated from the standard curve. In order to avoid the difference in cell amounts affecting the ATP measurements, the cellular proteins in each group were determined using the BCA assay kit and express ATP levels as nmol/mg protein.

For the ATP measurement of aqueous humor, 2–3µL of aqueous humor was carefully aspirate from the anterior chamber of each mouse eye using a fine-gauge needle, and after a cumulative total of 20 µL of aqueous humor was collected from each group, it was placed in centrifugation at 12,000 g for 10 min at 4 °C and the supernatant was collected. The ATP levels were subsequently determined and calculated.

GSH and GSSG determination

The intracellular levels of GSH and GSSG were measured using a GSH and GSSG assay kit. Briefly, TM cells were seeded in 6-well plates at a density of 1 × 10⁶ cells per well incubated overnight. Cells were exposed to H₂O₂ solution and treated with the exogenous mitochondria for 12 h (mitochondrial protein concentration: 2 µg proteins/mL). Cells were washed three times with ice-cold PBS, trypsinized and then centrifuged to collect pellets. 30 µL of protein removal reagent M solution was added to the cell pellets, followed by thorough vortex mixing. Subsequently, 20 µL of GSH removal aid solution was added into each 100ul of sample, vortexed and mixed immediately for subsequent analysis. The total glutathione and GSSG standard solutions were prepared according to the manufacturer’s instructions. Samples and standards were mixed with NADPH solution and incubated at room temperature for 25 min, the absorbance was measured at 412 nm using a micr1oplate reader. A standard curve was plotted for quantification. The content of GSH was calculated according to the following formula: GSH=Total Glutathione-GSSG×2.

Similarly, the GSSG levels in the samples were obtained according to the standard curve of GSSG, and the ratio of GSH/GSSG was calculated.

NAD⁺ and NADH determination

The intracellular levels of NAD + and NADH were determined according to the NAD⁺ and NADH assay kit with WST-8. Briefly, TM cells were seeded at a density of 1 × 10⁶ cells per well in 6-well plates and incubated overnight. Cells were exposed to H₂O₂ solution and treated with the exogenous mitochondria for 12 h (mitochondrial protein concentration: 2 µg proteins/mL). Cells were washed three times with PBS and 200 µL of NAD⁺/NADH extraction buffer was added per 1 × 10⁶ cells to lyse the cells. Subsequently centrifuged at 12,000 g, 4 °C for 10 min and the supernatant was taken as the sample. NADH standard solutions were prepared according to the manufacturer’s instructions. 90 µL of alcohol dehydrogenase working solution was added to both standards and samples, followed by 10-minute incubation at 37 °C in the dark to convert NAD⁺ to NADH. Then add 10 µL of colour development solution and incubate at 37 °C for 20 min to measure the absorbance at 450 nm on a microplate reader, and the standard curve was plotted to determine the total NAD + and NADH content (NAD _total). For NADH measurement, the samples should be lysed and heated at 60 °C for 30 min to degrade the NAD+. The intracellular NAD⁺ levels were calculated according to the following formula: NAD⁺= NAD_total- NADH.

Cell viability determination

Cell viability was detected by using the Annexin-VFITC/propidium iodide (PI) staining method. Briefly, cells were seeded in 6-well plates at a density of 1 × 10⁶ cells per well and cultured overnight. After treatment with different concentrations of H₂O₂ solution (1.0, 2.0 or 3.0 mM H₂O₂) for 12 h the cells were washed with PBS. For the mitochondria-treated group, isolated mitochondria were added and incubated for 12 h (mitochondrial protein concentration: 2 µg proteins/mL). Subsequently, cells were washed three times with ice-cold PBS, digested with trypsin, and collected. Cells were then resuspended in binding buffer (100 µL) and incubated with annexin V (5 µL) and PI staining solutions (10 µL) for 15 min at room temperature in dark. All samples were analyzed by flow cytometry within 30 min. The proportion of double-negative-stained cells represents the cell viability rate.

Western blot analysis

TM cells were lysed with RIPA lysis buffer containg a halt protease inhibitor cocktail. The lysate was centrifuged to collect proteins and the protein concentration was further quantified by a BCA assay kit. Separation gels and concentration gels were prepared and the proteins were added for electrophoresis.The proteins on the gels were transferred to polyvinylidene difluoride (PVDF) membrane. After 2 h of transfer, the PVDF membrane was cut and then washed three times with TBST. The primary antibodies of BAX, BCL-21, cleaved caspase-3, P62, Parkin, PINK-1, LC3B and GAPDH were diluted to a certain concentration and added to the membranes. After incubation at 4 °C overnight, the membranes were washed and treated with goat anti-rabbit IgG (H + L) or goat anti-mouse IgG (H + L) (Thermo Pierce, 31210, 1:1000) for 2 h at room temperature. Finally, washed with TBST three times and incubated with enhanced chemiluminescence kit (ECT) (Thermo Fisher Scientific, Waltham, USA) to visualise the bands on the membrane. Quantification was performed using Image J software. The expression of the proteins was normalised to the level of GAPDH.

3D cell spheroids assay

TM 3D cell spheroids were prepared using a modified hanging drop method [65, 66]. Briefly, TM cells were trypsinized and adjusted to a density of 5 × 10⁶ cells/mL, then 20 µL of cell suspension was pipetted onto the lid of a 24-well plate and the plate was inverted and placed in a CO₂ incubator for 48 h to allow spheroid formation.Cell spheroids with regular morphology and uniform size were taken for subsequent experiments.

In order to assess the ability and distribution of exogenous mitochondria into the interior of TM cell spheroids, the spheroids were treated with 2.0 mM H₂O₂ for 12 h, while the exogenous mitochondria were pre-labelled with Mito-Tracker Red (1µM, 37 °C, 10 min) and co-incubated with spheroids for 12 h (mitochondrial protein concentration: 2 µg proteins/mL). Spheroids were stained with Calcein AM and the nuclei were labeled by Hoechst. CLSM with Z-stack imaging was used to observe mitochondrial distribution across different cross-sections of the spheroids.

Crystal violet staining was used to observe the activity of cell spheroids before and after mitochondrial treatment. Firstly, TM cell spheroids were added to mitochondria after inducing damage with H₂O₂ (mitochondrial protein concentration: 2 µg proteins/mL). After 12 h, the cell spheroids were trypsinized into a single-cell suspension and stained with crystal violet. Dead cells were stained purple while live cells remain unstained. TM 3D cell spheroids survival rate was calculated as followed: Survival rate(%)= Unstained cells /(Stained cells +Unstained cells).

For the detection of ATP, GSH, GSSG, NAD⁺ and NADH, cell spheroids were first collected (10 spheroids per well), then dispersed into single-cell suspensions using trypsin (Fig. 5Aiii). Cells were collected after centrifugation. Subsequently, the same measurement method as described for the previous assays was followed.

OHT animal model and mitochondrial transplantation

To establish the OHT mice model, dexamethasone (DEX) was administered via topical ocular application. Since each eye responds differently to DEX, each eye can be treated as an independent data point. A total of 120 mice were randomly divided into 12 groups, with 20 eyes per group (n = 20). For the OHT model group, 0.1% DEX was applied to both eyes twice daily (once in the morning between 8 and 9 AM and once in the afternoon between 3 and 4 PM). The saline control group received topical saline solution in both eyes, while the healthy group without any treatment. After 22 consecutive days of treatment, when IOP stabilized, subsequent experiments were conducted. For the treatment groups, exogenous mitochondria (20 µg proteins/mL, 3µL) or an equivalent volume of PBS was injected into the anterior chamber of each eye (Fig. 7Ai-ii). ‌Timolol maleate eye drops‌ (0.5%, Wuhan Wujing Pharmaceutical Co., Ltd., China) and ‌latanoprost eye drops‌ (0.005%, Pfizer Inc., USA) were selected as positive controls for intraocular pressure-lowering drugs. Specifically, both anterior chamber injections‌ (PBS or mitochondria) ‌were single-dose administrations. Timolol maleate and latanoprost are recognized as first-line clinical therapeutics for IOP reduction, with a well-established efficacy benchmark [67, 68]. Therefore, these two agents were selected as positive controls in our study. ‌Timolol maleate eye drops were administered twice daily, while ‌latanoprost eye drops‌ were administered once daily. The specific groups were as follows:

• Healthy group without any treatment in both eyes (Healthy).

• Healthy with exogenous mitochondria injected in the anterior chamber (Healthy+Mito).

• Healthy with PBS injected in the anterior chamber (Healthy + PBS).

• Saline group with no treatment in the anterior chamber (Saline).

• Saline with exogenous mitochondria injected in the anterior chamber (Saline+Mito).

• Saline with PBS injected in the anterior chamber (Saline + PBS).

• DEX with no treatment in the anterior chamber (DEX).

• DEX with exogenous mitochondria injected in the anterior chamber (DEX+Mito).

• DEX with PBS injected in the anterior chamber (DEX + PBS).

• DEX with exogenous disrupted mitochondria injected in the anterior chamber (DEX+dMito).

• DEX with ‌Timolol maleate eye drops (DEX+Timolol).

• DEX with Latanoprost eye drops (DEX+Latanoprost).

IOP and weight measurements

IOP was measured using a rebound tonometer. Briefly, mice were anesthetized with 5% gaseous isoflurane (approximately 20 s), and IOP was measured in both eyes using the rebound tonometer (Icare TONOLAB; Vantaa, Finland) and conducted using a blinded methods. Each recorded IOP represents the average of 5 consecutive measurements and three sequential IOP readings were for the same eye to calculate the mean value. IOP measurements were performed every 2–3 days between 8:00–10:00 AM. To closely monitor fluctuations in IOP during the first week post-treatment (Days 23–30), IOP measurements were performed daily between 8:00–10:00 AM.

The mice were gently placed on an electronic balance (Electronic Scale, Kunshan, China) to measure their body weight. The body weight of each mouse was calculated as the average of three consecutive measurements. Body weight measurements were recorded every 2–3 days.

Distribution of exogenous mitochondria

The isolated mitochondria were stained with Mito-Tracker Green (1 µM, 37 °C, 10 min) and subsequently transplanted into the anterior chamber to investigate exogenous mitochondrial persistence. After injection, mice were anaesthetised at specified time points and the eyes were removed, fixed in 4% paraformaldehyde for 12 h, dehydrated in 30% sucrose and OCT-embedded and finally sectioned into frozen slices. Sections were stained with DAPI and the distribution of exogenous mitochondria was observed by CLSM. To calculate the distribution ratio of exogenous mitochondria within the TM, the following analytical procedure was employed. First, the total fluorescence intensity of the entire anterior chamber was measured using ImageJ software. Subsequently, the fluorescence intensity specifically localized to the TM region within the same field was quantified. The distribution ratio of exogenous mitochondria in the TM was then calculated by dividing the TM fluorescence intensity by the total anterior chamber fluorescence intensity.

Immunofluorescent staining

The cells or the frozen eye sections were permeabilized with 0.25% Triton X-100 and treated with bovine serum albumin for 30 min at room temperature. After blocking, the sections were incubated overnight with primary antibodies at 4 °C. Then, incubated for 2 h with fluorescent-labeled secondary antibody at room temperature. The nuclei were labeled by DAPI and analyzed under CLSM.

Proteomic analysis

For cell samples, 100 µl of SDS lysis buffer was added for cell lysis. After protein concentration determination using the BCA assay, sequential processing steps were performed, including trypsin digestion, protease desaltings, and analysis via liquid chromatography-tandem mass spectrometry (LC-MS/MS). Protein expression profiles were detected using data-independent acquisition (DIA) technology. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses (P < 0.05) were performed using their respective databases (http://geneontology.org/ and http://www.kegg.jp/) followed by visualization of the results.

ELIAS analysis

Enzyme-linked immunosorbent assay (ELISA) was used to determine the ROS and 8-hydroxydeoxyguanosine (8-OHdG) in the aqueous humor. Briefly, 20 µl aqueous humor samples were collected and followed by centrifugation at 12,000 rpm for 10 min to collect the supernatant. Then the ROS and 8-OHdG levels were determined according to the manufacturer’s instructions.

TEM observation of tissues’ mitochondria

At the time of harvest, mice were anaesthetised and eyes were removed and fixed in 2.5% glutaraldehyde 4 °C for 24 h. The eyes were then sectioned into 1 mm³ pieces, with care taken to avoid damaging the trabecular meshwork in each fragment. Subsequently, the pieces were post-fixed in 1% osmium tetroxide for 2 h at room temperature, dehydrated through an ethanol and acetone gradient, and embedded in resin. Resin blocks were sectioned into 1.5 μm semi-thin slices using a microtome. After toluidine blue staining, the trabecular meshwork was localized under a light microscope. And continued to be sectioned in 60–80 nm ultra-thin sections. These sections were sequentially stained with 2% uranyl acetate and 2.6% lead citrate in the dark, followed by observation and image acquisition under the TEM. Mitochondria morphology in the trabecular meshwork was analyzed in three randomly selected fields.

SEM observation of TM tissues’ structure

TM tissues were isolated following the methods described in previous studies [69]. Briefly, the eyes were cut into anterior and posterior halves, and the lens and vitreous body were gently removed. After posterior segment excision, the iris was gently grasped and pulled backward. The temporal side could be clearly identified as a pigmented stripe, representing the TM tissue. Surgical scissors were used to excise a 1 cm-wide tissue strip along both sides of the pigmented band. The specimens were immediately immersed in electron microscopy fixative, followed by dehydration and critical-point drying. For structural analysis, the samples were mounted with the TM surface exposed for SEM observation.

Aqueous humor clearance assessment

The aqueous humor clearance was measured following the methods described in previous studies with minor modifications [70]. Briefly, 1 mL of benzalkonium chloride (BAC, 0.05%) was instilled into the eyes of anesthetized mice to enhance corneal permeability to fluorescein. After 10 min, BAC was thoroughly washed with saline. Subsequently, 1 mL of 0.025% sodium fluorescein was applied to the eyes for 10 min, followed by saline washing. Fluorescence images were then captured every 10 min using a slit-lamp microscope under cobalt blue light (SLM-8E, China). The green fluorescence intensity was determined using ImageJ. Aqueous humor clearance was determined by calculating the decay constant of relative fluorescence intensity measured every 10 min over 60 min following a single fluorescein administration.

Safety and H&E evaluation

At the end of the experiment, the eyes of each group were examined using slit lamp.

To observe the structural changes in the trabecular meshwork post-treatment, hematoxylin and eosin (H&E) staining was performed. Briefly, the mice were anaesthetised, fixed in 4% paraformaldehyde, dehydrated, paraffin-embedded, sectioned, and stained with H&E. The trabecular meshwork structure was randomly observed using an Leica DM750 microscope.

To assess ocular surface safety analysis, at the treatment endpoint (55d), the mice were anesthetized and photographed the ocular surface under bright-field conditions using a slit lamp for observation.

To evaluate the systemic safety of mitochondrial treatment, major organs (heart, liver, spleen, lungs and kidneys) were excised and performed H&E staining.

Statistical analysis

All data generated were analyzed using SPSS 18.0 software. The experiments with in triplicate or more are expressed as mean ± standard deviation (SD). To compare differences between two groups, a two-sided Student’s t test was used. To determine differences between multiple groups, an ordinary one-way analysis of variance (ANOVA) was applied.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Xiaoling Wang and Yuanbo Liang designed the study (Conceptualization, Methodology). Xiaoling Wang, Beibei Lin, Xuegu Xu, Zhishu Bao, Peizhen Lin and Yongzhen Yu performed the experiments (Investigation). Xiaoling Wang, Beibei Lin, Guangying Luo and Ruiyi Ren analyzed the data (Formal analysis). Xiaoling Wang wrote the original draft (Writing- original draft). Xiaoling Wang, Yuanbo Liang, Zhishu Bao, Ruiyi Ren and Guangying Luo reviewed and edited the manuscript (Writing -review & editing). Xiaoling Wang, Yuanbo Liang, Guangying Luo and Ruiyi Ren visualized the results (Visualization). Yuanbo Liang and Xiaoling Wang supervised the project and provided funding (Supervision, Funding acquisition). All authors read and approved the final manuscript.

Funding

This work is supported by the National Natural Science Foundation of China (82070958), Zhejiang Provincial ‘Ten Thousand Talents Plan’ (2021R52012) and Natural Science Foundation of Zhejiang Province, China (LQ22H120004).

Data availability

The datasets supporting the findings of this study are included within the manuscript and its supplementary files. Additional data are available from the corresponding authors upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.