Modulation of Iron-Induced Glucose Metabolic Reprogramming Alleviates Retinal Pigment Epithelial Cell Senescence

Abstract

Purpose

This study aimed to investigate whether iron overload induces retinal pigment epithelial (RPE) cell senescence through glucose metabolic reprogramming and to evaluate the therapeutic potential of targeting this metabolic pathway.

Methods

We utilized human-induced pluripotent stem cell–derived RPE cells and induced RPE cells, along with mouse models with intravitreal or intraperitoneal injection of ferric ammonium citrate (FAC), to evaluate the effect of iron overload on RPE senescence. Proteomics, targeted metabolomics, reactive oxygen species (ROS) assay kits, JC-1 assay kit, reverse-transcription polymerase chain reaction, SA-β-gal staining, and western blot were used to assess mitochondrial function, ROS, and senescence markers. 2-Deoxy-d-glucose (2-DG), pyruvate kinase M2 inhibitor-1 (PKM2-IN-1), and sodium oxamate (SO) were used to modulate glucose metabolism flux. Flash electroretinography recording was used to assess visual function.

Results

Iron overload triggered significant glucose metabolic reprogramming in RPE cells, characterized by a time-dependent metabolic shift. Early exposure to FAC induced a transient surge in glucose metabolic flux, which elevated mitochondrial ROS production and disrupted mitochondrial homeostasis, ultimately leading to cellular senescence. Importantly, early inhibition of this metabolic surge with 2-DG or PKM2-IN-1 effectively attenuated senescence by reducing ROS levels and preserving mitochondrial function. Conversely, enhancing pyruvate flux with SO exacerbated senescence. The protective effect of 2-DG against iron-induced RPE senescence was further confirmed in a mouse model, where it preserved visual function and reduced senescence markers.

Conclusions

Glucose metabolic reprogramming mediates iron-induced RPE senescence, with transient glucose flux surge driving pathology via ROS-related mitochondrial damage. Targeting glucose metabolism may preserve mitochondria homeostasis and prevent RPE degeneration in age-related macular degeneration.

Age-related macular degeneration (AMD) is the most prevalent retinal disorder associated with aging and remains the leading cause of vision loss in individuals over 60 years of age.¹ Among the two clinical subtypes of AMD, the dry form accounts for approximately 90% of all cases and is characterized in its late stage by geographic atrophy of the macula and degeneration of the retinal pigment epithelial (RPE) cells.² The RPE cells play a critical role in maintaining retinal homeostasis by forming the outer blood–retina barrier, phagocytosing photoreceptor outer segments, and regulating the visual retinoid cycle.^3,4 Accumulating evidence has revealed that RPE cell senescence is the key etiology of AMD.^5–8 RPE cell senescence can be triggered by various intrinsic and extrinsic stressors, such as metabolic dysfunction, increased oxidative stress, and persistent inflammatory stimuli. Hallmarks of senescence include telomere shortening, persistent DNA damage responses, elevated senescence-associated β-galactosidase (SA-β-gal) activity, increased expression of cell cycle inhibitors such as p21, and the acquisition of a senescence-associated secretory phenotype (SASP).⁹

Iron is essential for numerous cellular processes, including adenosine triphosphate (ATP) production, DNA synthesis and repair, and oxygen transport, but its excess leads to oxidative stress, cell death, and senescence.^10–12 Age-related intracellular iron accumulation has been observed in multiple cell types and may contribute to disease progression.¹³ In the retina, iron accumulates in the macula of AMD patients, especially within the RPE and Bruch's membrane,¹⁴ and increases with aging.¹⁵ After anemic patients receive iron ion therapy, a large amount of lipofuscin, a typical feature of AMD, accumulates in the RPE cells.¹⁶ In induced pluripotent stem cell–derived RPE (iPSC-RPE) cells, chronic exposure to ferric ammonium citrate (FAC) leads to the accumulation of autofluorescent material, a hallmark of RPE aging and AMD.¹⁷ Furthermore, in ARPE-19 cells, iron released through ferritinophagy reduces mitochondrial membrane potential and impairs mitochondrial oxidative phosphorylation complexes, thereby promoting cellular senescence.¹⁸ Despite these findings, the mechanisms underlying iron-induced RPE cell senescence have not yet been fully elucidated. Clinically, there are currently no effective strategies to manage iron overload. Although iron chelators can lower iron levels, they are associated with significant side effects and RPE toxicity.¹⁹ Elucidating the mechanisms by which iron overload induces RPE cell senescence may therefore offer novel therapeutic targets for AMD treatment.

Mitochondria serve as central hubs for energy metabolism and redox homeostasis, producing ATP via oxidative phosphorylation and functioning as the primary site of reactive oxygen species (ROS) generation. Mitochondrial dysfunction is a recognized driver of cellular senescence, characterized by impaired biogenesis, ROS accumulation, and defective mitophagy.^20,21 Thus, maintaining mitochondrial homeostasis inhibits the senescence of RPE cells.^22,23 Additionally, mitochondria act as the central hub of glucose metabolism, where pyruvate is transported into the mitochondrial matrix to fuel the tricarboxylic acid (TCA) cycle.

Notably, excessive glucose metabolism is a hallmark of cellular senescence, and high-glucose conditions have been shown to induce senescence in various cell types.²⁴ Mechanistically, enhanced glucose metabolic flux or activation of pyruvate dehydrogenase induces an elevated mitochondrial pyruvate oxidation that promotes cell senescence by driving mitochondrial ROS production.^25–27 Mitochondria are also central organelles for iron utilization, with various iron-containing proteins playing key roles in mitochondrial energy metabolism.²¹ Therefore, we hypothesize that iron overload may alter glucose metabolism, thereby inducing mitochondrial homeostatic imbalance and RPE cell senescence. Modulating glucose metabolism is expected to maintain mitochondrial homeostasis and inhibit the iron overload–induced senescence of RPE cells.

In this study, we employed human iPSC-RPE cells and induced RPE (iRPE) cells, a lineage-stabilized cell line previously generated in our lab via transduction of key transcription factors into dedifferentiated iPSC-RPE cells,^28,29 as cell models. We aimed to investigate whether iron overload induces RPE cell senescence by triggering glucose metabolic reprogramming and damaging mitochondrial homeostasis. We further explored the therapeutic potential of modulating glucose metabolism to preserve mitochondrial function and attenuate RPE cell senescence. Ultimately, this work sought to identify novel therapeutic targets and strategies for the treatment of AMD.

Materials and Methods

Animal Experiments

This study utilized 7-week-old female C57BL/6 mice obtained from the Laboratory Animal Center of Tongji University in Shanghai, China. The mice were maintained under specific pathogen-free conditions with a controlled temperature of 25°C, relative humidity ranging from 40% to 70%, and a 12-hour light/dark cycle. All experimental animals were allowed free access to a standard laboratory rodent diet and drinking water. All experimental protocols were approved by the Animal Experimentation Ethics Committee of Tongji University (approval no. TJAA09625103) and conducted in compliance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research.

For intravitreal injections, mice were anesthetized with 5% Avertin, and 1 µL of FAC (100 µM) or saline (control) was injected weekly into the vitreous chamber using a 30-gauge needle. After 2 weeks, mice were euthanized for protein and RNA extraction or perfused with paraformaldehyde (PFA) for cryosection preparation. Embedded tissues were sectioned at 10 µm along the vertical meridian of the eye. For combination treatment, mice received weekly intravitreal injections of saline or FAC (100 µM, 1 µL), with or without intraperitoneal injections of 2-deoxy-d-glucose (2-DG; 1000 mg/kg every 3 days; TargetMol Chemicals, Boston, MA, USA) for 2 weeks. Flash electroretinography was performed at the end of treatment followed by euthanasia or PFA perfusion for RNA extraction or cryosection preparation. For systemic administration, mice received intraperitoneal saline or FAC (60 mg/kg) every 3 days for 3 months. At the end of treatment, mice were perfused with PFA, and the eyes, kidneys, and livers were collected for cryosection preparation.

Cell Cultures and Treatment Conditions

According to previous reports,²⁸ iPSC-RPE cells were cultured on Matrigel-coated dishes (NEST Biotechnology, Wuxi, China) in Dulbecco's Modified Eagle Medium/Nutrient Mixture F-12 (DMEM/F-12) supplemented with 10% knockout serum replacement (KSR), 50 U/mL penicillin, and 50 µg/mL streptomycin (all from Invitrogen, Carlsbad, CA, USA), and 1% non-essential amino acids and-10 mM nicotinamide (Sigma-Aldrich, St. Louis, MO, USA). iPSC-RPE cells were seeded at a density of 8 × 10⁴ cells/cm² and subjected to experimental treatments upon reaching approximately 80% confluence. Cells between passages P2 and P5 were used for all experiments. The iRPE cells were maintained in DMEM/F-12 containing 10% fetal bovine serum (FBS; Sino Biological, Beijing, China), 50 U/mL penicillin, and 50 µg/mL streptomycin. Cells were seeded at a density of 4 × 10⁴ cells/cm² and treated at approximately 80% confluence. iRPE cells between passages P10 and P20 were used.

The reagents used in this study included FAC (Hushi Chemical Company, Shanghai, China), 2-DG (TargetMol Chemicals), pyruvate kinase M2 inhibitor-1 (PKM2-IN-1; TargetMol Chemicals), and sodium oxamate (SO; TargetMol Chemicals). Unless otherwise indicated, cells were treated with FAC (200 µM), 2-DG (10 mM), PKM2-IN-1 (5 mM), or SO (10 mM or 30 mM) for the durations specified in each experiment. Prior to experimental treatments, iRPE cells were serum starved overnight in medium containing 1% FBS, whereas iPSC-RPE cells were treated directly in medium containing 10% KSR. To address the potential influence of serum conditions, iRPE cells were also cultured and assayed in medium supplemented with 10% FBS following the same FAC treatment protocol. All cell culture and treatment conditions were kept identical across experimental replicates to minimize variability.

Prussian Blue Staining Assay

Cellular iron deposition was assessed using the Prussian Blue Stain Kit (Yeasen Biotechnology, Shanghai, China) following the supplier's protocol. Following fixation with 4% PFA for 15 minutes, the cells or cryosections were rinsed three times with double-distilled water (ddH₂O). Equal volumes of Prussian blue solution A (potassium ferrocyanide) and solution B (acidic solution) were combined and applied to the samples. After a 15-minute incubation at room temperature, blue staining became evident. The samples were then washed three times with ddH₂O and counterstained with Nuclear Fast Red for 5 minutes. Imaging was performed under a light microscope (Olympus, Tokyo, Japan), and iron accumulation was quantitatively analyzed with ImageJ 1.53 (National Institutes of Health, Bethesda, MD, USA).

SA-β-gal Staining

SA-β-gal activity was assessed with a commercially available kit (C0602; Beyotime, Shanghai, China) in strict accordance with the provided guidelines. Sample fixation was carried out for 15 minutes at room temperature using the designated fixative solution from the kit. After three 5-minute phosphate-buffered saline (PBS) washes, a freshly prepared staining solution was applied. Incubation proceeded for 6 hours (iPSC-RPE cells) or 12 hours (iRPE cells and frozen sections) at 37°C under light-protected and carbon dioxide (CO₂)-free conditions. Image acquisition was conducted on an Olympus light microscope, with subsequent quantitative analysis of SA-β-gal–positive cells performed through ImageJ software.

For quantitative analysis of SA-β-gal staining in retinal cryosections, RPE regions were manually delineated in ImageJ based on predefined anatomical criteria. RGB images were split into individual color channels, and subsequent analysis was restricted to the RPE region. A blue/red ratio image was generated, converted to 8-bit grayscale, and thresholded using a fixed lower threshold determined from negative control sections. The same threshold was applied uniformly to all samples to generate binary masks. SA-β-gal positivity was quantified as the percentage of SA-β-gal–positive area relative to the total RPE area (area%).

EdU Cell Proliferation Assay

Cell proliferation was assessed using a BeyoClick EdU Cell Proliferation Kit with AF488 (C0071S; Beyotime) according to the manufacturer's instructions. Briefly, after treatment, cells were incubated with 5-ethynyl-2′-deoxyuridine (EdU) working solution (10 µM final concentration) at 37°C for 4 hours to allow incorporation of the EdU into newly synthesized DNA. After incubation, cells were washed with PBS and fixed with 4% paraformaldehyde for 15 minutes at room temperature. Cells were then permeabilized with 0.3% Triton X-100 (Sigma-Aldrich) in PBS for 15 minutes and washed thoroughly. The incorporated EdU was detected using the provided click-reaction cocktail in the dark at room temperature for 30 minutes. Cell nuclei were counterstained with Hoechst 33342 (C1027; Beyotime). Fluorescence images were captured using an Olympus fluorescence microscope with appropriate filter settings. All images were acquired using identical exposure parameters across experimental groups. EdU-positive cells were quantified and expressed as a percentage of total nuclei.

Autofluorescence Imaging

Cellular autofluorescence was assessed in live iPSC-RPE cells without exogenous fluorescent probes. Following treatment, cells were rinsed once with PBS and then incubated in phenol red–free medium. Autofluorescence imaging was performed using an Olympus light microscope with a 488-nm excitation wavelength; crucially, identical exposure times and gain parameters were applied across all experimental groups to guarantee consistency. To minimize photobleaching and maximize data comparability, all images were captured immediately after sample preparation.

Calcein-AM/Propidium Iodide Staining

Cell viability was assessed using Calcein-AM/propidium iodide (PI) dual staining with a commercial kit (Beyotime). Cells were seeded in 96-well plates and subjected to the indicated treatments. After treatment, cells were washed twice with PBS and incubated with 100 µL of Calcein-AM/PI working solution (prepared according to the manufacturer's instructions) at 37°C for 30 to 60 minutes in the dark. Following incubation, cells were gently washed with PBS, and fluorescent images were acquired using an Olympus fluorescence microscope. The ratio of dead cells was quantified as the number of PI-positive cells/(number of Calcein-positive cells + number of PI-positive cells).

Detection of Intracellular ROS with DCFH-DA

Total intracellular ROS levels were measured using 2′,7′-dichlorofluorescein diacetate (DCFH-DA) from a ROS Assay Kit (S0035; Beyotime). After treatment, cells were washed twice with PBS and incubated with 10-µM DCFH-DA (final concentration) in PBS at 37°C for 10 minutes in the dark. Excess probe was removed by washing cells three times with PBS. Fluorescence intensity was measured using a Synergy H1 microplate reader (BioTek Instruments, Winooski, VT, USA) with an excitation wavelength of 488 nm and emission wavelength of 525 nm. For imaging-based analysis, fluorescence was visualized using an Olympus fluorescence microscope with strictly identical imaging settings across all groups, including the same excitation wavelength and camera exposure time during image acquisition.

Mitochondrial Superoxide Detection With MitoSOX Red

Mitochondrial superoxide production was assessed using MitoSOX Red (S0061; Beyotime). Following treatment, cells were washed twice with warm PBS and incubated with 5-µM MitoSOX Red (final concentration) diluted in PBS at 37°C for 10 minutes in the dark. For nuclear counterstaining, Hoechst 33342 was added simultaneously at a dilution of 1:100. Cells were then washed three times with PBS to remove excess probe. Fluorescence images were captured using an Olympus fluorescence microscope with appropriate filter settings. All images were acquired using the same exposure and gain settings. MitoSOX fluorescence intensity was used to determine the mitochondrial superoxide levels.

Mitochondrial Membrane Potential Detection

Mitochondrial membrane potential was assessed with a JC-1 Assay Kit (C2006; Beyotime). Following FAC treatment, a single PBS wash was performed. The JC-1 reagent was prepared at 10 µM in DMEM/F-12 medium, applied to the cells, and incubated for 20 minutes. After removal of the staining solution, the cells were rinsed once with PBS. Fresh medium was introduced prior to fluorescence observation. Images of both red and green fluorescence were acquired using a fluorescence microscope. Quantitative fluorometric analysis was carried out by detecting JC-1 monomers at 490-nm excitation/530-nm emission and J-aggregates at 525-nm excitation/590-nm emission, with fluorescence intensity evaluated using ImageJ software.

NAD⁺/NADH Measurement

Intracellular nicotinamide adenine dinucleotide (NAD)-positive and nicotinamide adenine dinucleotide hydride (NADH) concentrations were determined with a commercial NAD⁺/NADH Assay Kit (S0175; Beyotime) in accordance with the manufacturer's protocol. Briefly, cells (1 × 10⁶ per sample) were harvested and lysed using the lysis buffer provided in the kit. To assess total NAD⁺/NADH (NADtotal), 20 µL of cytosolic lysate was aliquoted into a 96-well plate. For NADH measurement, an equivalent volume of lysate was first heated at 60°C for 30 minutes. Thereafter, 90 µL of alcohol dehydrogenase solution was added to each well and the plate was incubated at 37°C for 10 minutes. Next, 10 µL of chromogen was introduced to each well and allowed to react for another 10 minutes at 37°C. Absorbance was read at 450 nm using a microplate reader (Bio-Rad, Hercules, CA, USA) alongside a standard curve. NAD⁺ content was derived by subtracting NADH from NADtotal, and the NAD⁺/NADH ratio was computed accordingly.

ATP Production

Intracellular ATP levels were detected using an ATP Assay Kit (S0027; Beyotime). The operation was carried out according to the instructions. Protein concentrations were quantified by the bicinchoninic acid (BCA) method. The chemiluminescence signals were measured by a chemiluminescence detection system at a 0.2-second exposure time and normalized to protein concentrations.

Pyruvate Dehydrogenase Activity Assay

Pyruvate dehydrogenase (PDH) activity was measured using a commercial PDH Activity Assay Kit (BC0380; Solarbio, Beijing, China) according to the manufacturer's instructions. Cells treated with FAC for different durations were washed once with ice-cold PBS, collected by scraping on ice, and immediately processed for PDH activity measurement. After protein quantification, PDH activity was normalized to total protein content, and the normalized values were used for subsequent statistical analysis.

Cell Viability Assay

Cell viability was determined using a Cell Counting Kit-8 (C0005; TargetMol Chemicals). The iRPE cells were seeded in 96-well plates at 1.0 × 10⁴ cells per well and subjected to experimental treatments. Subsequently, CCK-8 working solution was applied to all wells, and the plates were maintained at 37°C for 2 hours. Optical density was then measured at 450 nm with an iMark Microplate Absorbance Reader (Bio-Rad).

Quantitative Real-Time Polymerase Chain Reaction

Total RNA was extracted from cells, and cDNA synthesis was carried out with the Primescript RT Master Mix kit (Takara, Shiga, Japan). Quantitative real-time polymerase chain reaction (qRT-PCR) was performed on a Chromo4 system (Bio-Rad) employing Superreal Premix Plus reagent (Tiangen Biotech, Beijing, China). The thermal cycling conditions included an initial denaturation step at 95°C for 5 minutes, followed by 40 cycles of 95°C for 30 seconds and 60°C for 30 seconds. All primers, provided in Supplementary Table S2, were synthesized by Sangon Biotech (Shanghai, China).

Western Blot

Protein extraction was carried out with radioimmunoprecipitation assay (RIPA) buffer containing protease and phosphatase inhibitors (Sigma-Aldrich). A total of 20 µg protein per sample was separated via 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and transferred electrophoretically to polyvinylidene fluoride (PVDF) membranes (MilliporeSigma, Bedford, MA, USA). Membranes were blocked for 1 hour with 5% bovine serum albumin (BSA) prepared in Tris-buffered saline with Tween 20 (TBST), then probed with primary antibodies at 4°C overnight. After washing, membranes were incubated with corresponding horseradish peroxide (HRP)-conjugated secondary antibodies for 1 hour at room temperature. Detection was performed using a chemiluminescence imaging system (5200; Tanon, Shanghai, China), and band intensities were quantified with ImageJ software. Antibody information is available in Supplementary Table S1.

Immunostaining

Immunofluorescence staining was performed on cultured cells fixed with 4% PFA for 15 minutes and rinsed three times with PBS. The cells were then permeabilized with 0.1% Triton X-100 for 10 minutes, followed by PBS rinsing and blocking of nonspecific sites with 3% BSA in PBS for 1 hour at room temperature. Samples were then incubated overnight at 4°C with primary antibodies directed against γ-H2A.X. Following three washes with PBS, the cells were incubated with fluorescently conjugated secondary antibodies (Invitrogen) for 1 hour at room temperature. Nuclei were counterstained using 4′,6-diamidino-2-phenylindole dihydrochloride (DAPI; Sigma-Aldrich). Imaging was performed using an Olympus fluorescence microscope. A comprehensive list of antibodies used is provided in Supplementary Table S1.

Fe²⁺ Measurement

To assess intracellular Fe²⁺ levels, cells seeded in a 48-well plate were treated with 200-µM FAC for the indicated durations. After treatment, the medium was replaced with DMEM/F-12 containing 1-µM FerroOrange (DOJINDO Laboratories, Kumamoto, Japan), and cells were incubated at 37°C under 5% CO₂ for 30 minutes. Following staining, fluorescence images were acquired using an Olympus IX73 microscope. The relative Fe²⁺ level was quantified as the mean fluorescence intensity per cell.

Flash Electroretinogram Recording

Flash electroretinography was conducted with an APS-2000 visual electrophysiology system (Kanghua Ruiming Science Technology, Chongqing, China). After overnight (12-hour) dark adaptation, animals were anesthetized and positioned under dim red light. Corneal contact electrodes were placed along with subdermal reference and ground electrodes along the midline. ERG responses were recorded at a flash intensity of 6.325 × 10^–2 cd·s/m².

Protein Profiling Analysis

Following cell lysis, protein concentration was measured with a BCA assay. Aliquots containing 50 µg of protein from each sample underwent tryptic digestion and were subsequently labeled with tandem mass tag (TMT) reagents. Peptides were separated using an Agilent 1100 HPLC System with a ZORBAX Extend-C18 column (2.1 mm × 150 mm, 5 µm) (Agilent Technologies, Santa Clara, CA, USA) and a binary gradient consisting of mobile phase A (2% acetonitrile/water) and B (90% acetonitrile/water) at a flow rate of 300 µL/min, with detection at 210 nm. The separated fractions were lyophilized before mass spectrometry analysis. Mass spectrometric data were acquired on a Q-Exactive HF instrument (Thermo Fisher Scientific, Waltham, MA, USA) equipped with a Nanospray Flex Ion Source and coupled to an EASY-nLC 1200 system (Thermo Fisher Scientific). Peptides were separated on a C18 column (75 µm × 50 cm). Full mass spectrometry (MS) scans were performed over 350 to 1500 m/z at a resolution of 60,000, with an automatic gain control (AGC) target of 3 × 10⁶. The top 20 most intense precursors were selected for fragmentation via higher-energy collisional dissociation (HCD) at 32% normalized collision energy. Tandem mass spectrometry (MS/MS) spectra were acquired at a resolution of 45,000, with an AGC target of 2 × 10⁵ and a maximum injection time of 80 ms. Dynamic exclusion was set to 30.0 seconds. Quantitative proteomic analysis was conducted by Oebiotech (Shanghai, China). The MS proteomics data have been deposited to the ProteomeXchange Consortium (https://proteomecentral.proteomexchange.org) via the iProX partner repository with the dataset identifier PXD072937.

Bioinformatic and Enrichment Analysis

Differentially expressed proteins were subjected to gene set enrichment analysis (GSEA) using GSEA software (version 4.3.2). Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis was performed using KEGG Mapper (KEGG database release 109.0). Pathways with adjusted P < 0.05 were considered significantly enriched.

Targeted Energy Metabolomics With LC-MS/MS

Energy metabolites were quantified by targeted liquid chromatography/tandem mass spectrometry (LC-MS/MS) using a Waters ACQUITY UPLC H-Class System (Waters, Milford, MA, USA) coupled to a QTRAP 6500+ system (SCIEX, Framingham, MA, USA). Cell pellets were resuspended in 100 µL ultrapure water, mixed with 200 µL precooled methanol, and subjected to three freeze/thaw cycles (liquid nitrogen/ice). Samples were centrifuged at 12,000 rpm for 10 minutes at 4°C, and 180 µL of supernatant was collected for analysis. Separation was performed on an ACQUITY UPLC BEH Amide Column (2.1 × 100 mm, 1.7 µm; Waters) with a gradient from 95% to 50% acetonitrile (with 10-mM ammonium acetate and 0.3% ammonium hydroxide in water) over 11 minutes, returning to 95% by 15 minutes at 0.4 mL/min and 40°C. Metabolites were detected using electrospray ionization (ESI) in positive and negative modes with the Scheduled MRM algorithm, and quantified using Analyst 1.6.3 and MultiQuant 3.0.3 software (SCIEX) with optimized declustering potentials and collision energies. Finally, the supernatants were collected for LC-MS/MS analysis. All of the metobolites were detected by MetWare (Wuhan, China) based on the the SCIEX QTRAP 6500 LC-MS/MS platform. Heatmaps were generated using MetaboAnalyst 5.0. No statistical filtering or preselection was applied prior to visualization, and all quantified metabolites were included in the heatmap analysis.

Statistical Analysis

All statistical analyses were performed using Prism 10 (GraphPad, Boston, MA, USA). Data are presented as mean ± standard deviation (SD). For group comparisons with small sample sizes (n ≤ 4), nonparametric tests were applied, including the Mann–Whitney U test for two-group comparisons and the Kruskal–Wallis test for multiple-group comparisons. Notably, when n = 3, the minimum attainable P value was 0.10; therefore, all comparisons with n = 3 are annotated as P = 0.10 in the figures.

For group comparisons with larger sample sizes (n ≥ 5), parametric tests were used, including unpaired, two-tailed Student's t-tests for two-group comparisons or one-way analysis of variance (ANOVA) tests for multiple comparisons. P < 0.05 was considered statistically significant. For group comparisons with a sample size of n ≥ 5, the normality of data distribution was first assessed using the Shapiro–Wilk test. Data passing the normality test (P > 0.05) were analyzed using parametric tests: unpaired, two-tailed Student's t-tests for two-group comparisons or one-way ANOVA for multiple comparisons. If the normality assumption was violated, non-parametric tests were applied. For comparisons with small sample sizes (n ≤ 4), nonparametric tests (Mann–Whitney U test for two groups; Kruskal–Wallis test for multiple groups) were used directly. P < 0.05 was considered statistically significant. For comparisons with n = 3, a P value of 0.10 is annotated in the figures whenever it was attained, as this represents the minimum attainable P value at this sample size.

Results

Iron Overload Triggers RPE Cell Senescence

To confirm that iron overload induces RPE cell senescence, we first established both local and systemic iron overload mouse models via intravitreal or intraperitoneal injection of FAC, respectively. In mice receiving intravitreal FAC, RPE cells displayed strong SA-β-gal–positive staining (Fig. 1A), and the p21 protein level was elevated in the RPE–choroid complex compared to that in the control (Fig. 1B). Additionally, SASP genes, including IL6, CCL2, and CXCL11, were upregulated in the RPE–choroid complex (Fig. 1C). These results confirm that iron overload induces RPE cell senescence. Furthermore, in the systemic iron overload model, FAC-treated mice exhibited robust iron accumulation in the RPE cells and increased SA-β-gal staining (Supplementary Figs. S1A–S1C). Moreover, we observed senescence in non-RPE tissues, including the kidney and liver, under conditions of iron overload (Supplementary Figs. S1D, S1E), suggesting that iron-induced senescence represents a widespread cellular response.

Figure 1.

Iron overload induces RPE cell senescence in vivo. Mice received weekly intravitreal injections of either saline or FAC (100 µM, 1 µL) for 2 weeks. (A) SA-β-gal staining of mouse retinal sections and quantitative analysis of the percentage of SA-β-gal–positive area within the RPE layer (n = 4). (B) Western blot analysis of p21 protein levels in mouse RPE–choroid complexes (n = 3). (C) mRNA expression levels of SASP genes in the mouse RPE–choroid complex were assessed by qRT-PCR (n = 3). Scale bar: 20 µm. Data are expressed as mean ± SD. Statistical significance was determined by Mann–Whitney U tests. *P < 0.05.

In vitro, we employed two RPE cell types: iPSC-RPE cells and iRPE cells to confirm that iron overload induces RPE cell senescence. Long-term exposure to 200-µM FAC (10 days for iPSC-RPE, 3 days for iRPE) led to marked intracellular iron accumulation (Supplementary Figs. S2A, S2B), and both cell types showed a markedly increased number of SA-β-gal–positive cells (Figs. 2A, 2B), consistent with a senescent phenotype. The qRT-PCR analysis further confirmed robustly upregulated expression of SASP genes: In iPSC-RPE cells, IL-1A, IL6, CCL2, CXCL1, CXCL11, and SERPINE1 were upregulated, whereas, in iRPE cells, IL6, CCL2, CXCL1, and SERPINE1 showed increases (Figs. 2C, 2D). In addition, as evidenced by EdU staining (Figs. 2E, 2F), both iPSC-RPE and iRPE cells exhibited a significant decrease in cell proliferation. Western blot analysis further showed an upregulation of the cell cycle arrest markers p16 and p21 in both cell types after iron treatment (Figs. 2G, 2H). Consistent with these findings, CCK-8 assays revealed a reduction in absorbance in both cell lines upon iron treatment, consistent with decreased proliferative capacity. (Fig. 2I). In iPSC-RPE cells, iron overload also induced prominent autofluorescence (Figs. 2J, 2K), a hallmark of RPE cell senescence.³⁰ Importantly, iron overload–induced senescence appeared to be time dependent, as SA-β-gal staining following 5-day FAC treatment in iPSC-RPE cells and 1-day treatment in iRPE cells showed no significant changes compared to controls (Supplementary Figs. S2C–S2F), suggesting that a certain threshold of iron accumulation or exposure duration is required to trigger RPE cell senescence. Together, these results demonstrate that chronic iron overload induces senescence in both types of RPE cells. Notably, this iron-induced senescence phenotype was also confirmed in iRPE cells cultured under serum-containing conditions (Supplementary Figs. S3), indicating the robustness of this response across different culture paradigms. Because FAC treatment induces both iron overload and elevated ammonium citrate levels, we treated iPSC-RPE cells with ammonium citrate alone at equivalent concentrations and duration. SA-β-gal staining revealed no significant change in these control cells (Supplementary Figs. S4A, S4B), indicating that iron overload, rather than ammonium citrate, is the primary driver of FAC-induced RPE cell senescence.

Figure 2.

Iron overload induces RPE cell senescence in vitro. iPSC-RPE and iRPE cells were treated with or without FAC (200 µM) for 10 days and 3 days, respectively. (A) SA-β-gal staining images. (B) Quantitative analysis of the percentage of SA-β-gal–positive cells in A (n = 5). (C, D) mRNA expression levels of SASP genes in iPSC-RPE and iRPE cells were analyzed by qRT-PCR (n = 3). (E, F) Proliferation was assessed by EdU staining in iPSC-RPE and iRPE cells and quantified as the ratio of EdU⁺ cells (n = 4). (G) Western blot analysis of p21, p16, and γ-H2A.X expression in iPSC-RPE cells (n = 3). (H) Western blot of p21 and p16 in iRPE cells (n = 3). (I) CCK-8 viability assay of iPSC-RPE and iRPE cells (n = 4). (J, K) Autofluorescence of FAC-treated iPSC-RPE cells and quantitative analysis of the intensity of autofluorescence (n = 5). Scale bar: 20 µm. Data are expressed as mean ± SD. Statistical significance was determined by Mann–Whitney U tests. *P < 0.05; **P < 0.01.

Given that iron overload can also trigger cell death, such as ferroptosis, we monitored cell viability over time using Calcein-AM/PI staining. When treated with 200-µM FAC, no overt cell death was detected at the examined time points (day 5 and day 10 for iPSC-RPE cells; day 1 and day 3 for iRPE cells), as indicated by the minimal PI-positive staining observed, suggesting limited cell death (Supplementary Figs. S5A–S5D). However, prolonged FAC exposure beyond these time points (day 20 for iPSC-RPE cells; day 6 for iRPE cells) ultimately induced cell death in both cell types (Supplementary Figs. S5E–S5G). This observation is consistent with previous reports that iron overload can trigger ferroptosis in RPE cells.²⁸ Moreover, although FAC treatment induced modest mRNA-level changes in some ferroptosis-associated genes (e.g., ACSL4, SLC7A11) in iPSC-RPE (day 10) and iRPE (day 3) cells, the expression of their corresponding proteins remained unaltered (Supplementary Figs. S5H–S5J). By contrast, the iron-storage proteins FTH1 and FTL were markedly upregulated (Supplementary Fig. S5H). Together, these data indicate that, under these experimental conditions, iron overload primarily induces intracellular iron accumulation and cellular senescence in RPE cells prior to the onset of significant cell death.

Iron Overload Elevates ROS and Disrupts Mitochondrial Homeostasis in RPE Cells

To elucidate the mechanisms underlying iron overload–induced RPE cell senescence, we conducted proteomic profiling of iPSC-RPE cells and demonstrated the extensive protein expression changes between the control and FAC-treated groups (Fig. 3A). GSEA highlighted significant downregulation of mitochondrial function–related pathways, including the mitochondrial inner membrane, oxidative phosphorylation, and mitochondrial matrix (Fig. 3B). These results suggest that iron overload may alter mitochondrial function and homeostasis. To further clarify the effect of iron overload on mitochondrial function, we performed energy metabolomic analysis of iPSC-RPE cells, which revealed marked differences between the control and FAC-treated samples, as illustrated by the heatmap (Fig. 3C). Notably, ATP levels in the FAC group were reduced by more than twofold (Fig. 3D). This decrease was confirmed in both iPSC-RPE and iRPE cells (Fig. 3E), accompanied by a significant reduction in the NAD⁺/NADH ratio (Fig. 3F) after FAC treatment. Additionally, both iPSC-RPE and iRPE cells treated with FAC exhibited a significant reduction in mitochondrial membrane potential (Figs. 3G–J). Colocalization analysis using FerroOrange and MitoTracker further revealed an accumulation of iron within mitochondria at the mid-stage of FAC treatment (5 days for iPSC-RPE cells and 2 days for iRPE cells), which was quantitatively confirmed by a significant increase in FerroOrange fluorescence intensity per cell (Supplementary Figs. S6A, S6B). To investigate the mechanisms underlying iron-induced mitochondrial homeostatic imbalance, we assessed the levels of ROS. Notably, mitochondrial superoxide (detected by MitoSOX) was markedly elevated at these intermediate time points (Figs. 3K–M), and overall cellular ROS levels (measured by DCFH-DA) were also increased (Figs. 3N, 3O). Together, these results indicate that sustained iron overload elevates mitochondrial and cellular ROS, impairs mitochondrial function, and contributes to iron-induced senescence in RPE cells.

Figure 3.

Iron overload disrupts mitochondrial homeostasis in RPE cells. MS was used to quantify the differentially expressed proteins (DEPs) in iPSC-RPE cells treated with or without FAC (200 µM) for 10 days. (A) Volcano plot of DEPs in iPSC-RPE cells after FAC treatment; 540 proteins were upregulated and 245 proteins were downregulated. (B) GSEA plots showing downregulated mitochondrial inner membrane, mitochondrial matrix, and oxidative phosphorylation pathways. (C, D) The metabolites of glycolysis and TCA in iPSC-RPE cells after treatment with or without FAC for 10 days were assayed by MS. Heatmap shows Z-scored relative abundances of all detected metabolites (C); also shown is a quantitative analysis of ATP (n = 3) (D). (E–J) iPSC-RPE and iRPE cells treated with or without FAC for 10 days and 3 days, respectively. ATP levels were analyzed by an ATP assay kit (n = 4) (E). NAD⁺/NADH ratios (n = 5) (F). JC-1 staining and fluorescence quantification of mitochondrial membrane potential (n = 5) (G–J). (K–O) iPSC-RPE cells and iRPE cells were treated with or without FAC (200 µM) for 5 days and 2 days, respectively. MitoSOX Red staining and quantitative analysis of fluorescence density were performed (n = 4) (K–M). ROS levels were assessed by a DCFH-DA fluorescence assay kit (n = 4) (N, O). Scale bar: 20 µm. Data are expressed as mean ± SD. Statistical significance was determined by Mann–Whitney U tests (D–F, L–O) or unpaired two-tailed Student's t-tests (I, J). *P < 0.05; ***P < 0.001.

Iron Overload Reprograms Glucose Metabolism and Induces a Time-Dependent Metabolic Shift in RPE Cells.

To investigate if iron overload impacts glucose metabolic processes, we first analyzed differentially expressed proteins in FAC-treated iPSC-RPE cells. KEGG pathway analysis revealed significant enrichment of glycolysis-related pathways (Fig. 4A), whereas protein–protein interaction (PPI) analysis identified upregulation of key glycolytic enzymes including ENO1, GAPDH, LDHA, PGK1, and PKM among the top 25 connectivity rankings (Fig. 4B), suggesting substantial alterations in glucose metabolism following iron overload. In particular, analysis of glucose metabolism in iPSC-RPE cells after 10 days of FAC treatment revealed a distinct metabolomic profile compared to controls (Fig. 3C). Notably, cis-aconitate, a TCA intermediate, was elevated, whereas acetyl-CoA was decreased following FAC treatment (Figs. 4C, 4D). Additionally, intracellular glucose accumulation was observed (Fig. 4E). To investigate how iron overload impairs the TCA cycle and glucose metabolic flux, we first analyzed pyruvate, the critical link between glycolysis and the TCA cycle. In iPSC-RPE cells, pyruvate levels increased at day 1 post-FAC treatment, returned to baseline by day 4, and declined by day 10 (Fig. 4F). A similar trend was observed in iRPE cells, with levels rising at day 0.5, normalizing by day 2, and subsequently declining by day 3 (Fig. 4G). Correspondingly, PDH activity was transiently upregulated early after treatment (peaking at day 1 in iPSC-RPE and day 0.5 in iRPE cells) but became significantly suppressed in iPSC-RPE cells by days 4 and 10 and declined to near-baseline levels in iRPE cells by day 3 (Figs. 4H, 4I). This transient activation pattern was also reflected in ATP levels, which showed an early increase followed by a decrease in both cell types (Fig. 4J). Finally, CCK-8 assays indicated that the reduction in cell viability occurred only at later stages, consistent with the delayed suppression of metabolic activity (Supplementary Figs. S6B, S6C). Collectively, these findings suggest a time-dependent metabolic shift in FAC-treated RPE cells, in which early FAC exposure is associated with a transient increase in glucose metabolic flux, whereas prolonged treatment is accompanied by features consistent with TCA cycle impairment and reduced energy metabolism.

Figure 4.

Iron overload drives glucose metabolism reprogramming in RPE cells. MS was used to quantify the DEPs in iPSC-RPE cells treated with or without FAC (200 µM) for 10 days. (A) KEGG pathway enrichment highlighted that the glycolysis-related pathway (red) was altered by FAC treatment in iPSC-RPE cells. (B) PPI analysis was performed with DEPs; proteins with the top 25 connectivity rankings were selected for re-conducting PPI analysis. Among these proteins, those related to glucose metabolism, including ENO1, GAPDH, LDHA, PGK1, and PKM, were highly connected and upregulated. (C–E) The metabolites of glucose in iPSC-RPE cells after being treated with or without FAC for 10 days were assayed by MS and quantitative analysis of cis-aconitate, acetyl-CoA, and D(+)-glucose (n = 3). (F–J) Analysis of metabolic parameters in iPSC-RPE and iRPE cells following treatment with 200-µM FAC for the indicated durations (n = 4 for F–I; n = 5 for J). Pyruvate levels (F, G). Relative PDH activity (H, I). ATP levels (J). Data are expressed as mean ± SD. Statistical significance was determined by Mann–Whitney U tests (C–E), Kruskal–Wallis tests (F–I), or one-way ANOVA (J). *P < 0.05; ***P < 0.001; ns, not significant (P ≥ 0.05).

Glucose Metabolic Surge Is Associated With Iron-Induced RPE Senescence and Mitochondrial Stress

Based on the observed time-dependent metabolic shift, we hypothesized that the transient increase in glucose metabolic flux following FAC exposure may increase mitochondrial workload, thereby exacerbating iron-dependent ROS generation and favoring the emergence of senescence-like phenotypes. To examine this possibility, we pharmacologically modulated glucose metabolism to either attenuate or sustain the early metabolic surge and assessed its association with oxidative stress, mitochondrial function, and senescence markers. We first inhibited glucose metabolic flux with the hexokinase blocker 2-DG (10 mM), which suppresses glycolysis and subsequent pyruvate production.³¹ Early 2-DG treatment (from day 0 to day 1 in iPSC-RPE or to day 0.5 in iRPE cells) effectively suppressed the FAC-induced pyruvate surge, indicating an attenuation of glucose metabolic flux (Figs. 5A, 5B). Notably, only early and sustained 2-DG administration (from day 0 until endpoint) attenuated iron-induced senescence (SA-β-gal staining) (Figs. 5C–E), whereas delayed treatment (started on day 5 in iPSC-RPE or day 1.5 in iRPE cells) offered no protection (Supplementary Figs. S7A–S7C). Although 2-DG did not affect iron accumulation (Supplementary Figs. S7D, S7E), it significantly lowered ROS levels (measured at day 5 in iPSC-RPE and day 2 in iRPE cells) (Figs. 5F, 5G) and preserved mitochondrial membrane potential (assessed at day 10 in iPSC-RPE and day 3 in iRPE cells) (Figs. 5H–K). These results indicate that blocking the glucose metabolic flux alleviates senescence mainly by reducing oxidative stress and maintaining mitochondrial function. To further validate this mechanism, we employed PKM2-IN-1 (5 µM, from day 0 to endpoint), a pyruvate kinase M2 inhibitor–reducing glucose metabolic flux.³² Consistent with 2-DG, PKM2 inhibition reduced both senescence and ROS levels in FAC-treated iPSC-RPE cells (Figs. 5L–Q). Together, these data demonstrate that the transient surge in glucose metabolic flux plays a key role in driving iron-induced RPE senescence.

Figure 5.

Inhibiting glucose metabolic flux mitigates iron-induced RPE cell senescence. (A, B) Pyruvate levels in iPSC-RPE cells (n = 4) and iRPE cells (n = 4) treated with FAC (200 µM) ± 2-DG (10 mM) for 1 day and 0.5 day, respectively. (C–K) iPSC-RPE and iRPE cells were treated with or without FAC (200 µM) ± 2-DG (10 mM) for 10 days and 3 days, respectively. SA-β-gal staining (C). Quantitative analysis of the percentage of SA-β-gal⁺ cells in C (n = 5) (D, E). DCFH-DA staining and ROS quantification analysis (n = 4) (F, G). JC-1 staining and mitochondrial membrane potential quantification analysis (n = 4) (H–K). (L–Q) iPSC-RPE cells were treated with FAC (200 µM) ± PKM2-IN-1 (5 mM). SA-β-gal staining of iPSC-RPE cells after being treated for 10 days and quantitative analysis of the percent of SA-β-gal⁺ cells (n = 4) (L, M). DCFH-DA staining and ROS quantification in iPSC-RPE cells after being treated for 5 days (n = 4) (N, O). MitoSOX Red staining and quantitative analysis of mean fluorescence intensity (MFI) in iPSC-RPE cells after being treated for 5 days (n = 4) (P, Q). Scale bar: 20 µm. Data are expressed as mean ± SD. Statistical significance was determined by Kruskal–Wallis tests (A, B), unpaired two-tailed Student's t-tests (D, E), or Mann–Whitney U tests (I, M, K, O, Q). *P < 0.05; ***P < 0.001; ns, not significant (P ≥ 0.05).

Conversely, the lactate dehydrogenase A (LDHA) inhibitor SO (10 mM) blocked pyruvate-to-lactate conversion and caused pyruvate accumulation (Fig. 6A). In iPSC-RPE cells treated with SO for the full duration of FAC exposure (from day 0 to day 10), senescence markers and ROS levels increased (SA-β-gal, γ-H2A.X) (Figs. 6B–F). Moreover, high-dose SO (30 mM) treatment for 10 days alone was sufficient to elicit senescence-like phenotypes (Figs. 6G, 6H), suggesting that persistent metabolic burden may contribute to senescence independently of iron overload.

Figure 6.

SO treatment promotes RPE senescence. (A) Pyruvate levels in iPSC-RPE cells treated for 2 days (SO, 10 mM) were analyzed by an assay kit (n = 4). (B–F) iPSC-RPE cells were treated with or without FAC (200 µM) ± SO (10 mM) for 10 days. SA-β-gal staining and quantitative analysis of the percentage of SA-β-gal+ cells (n = 4) (B, C). γ-H2A.X immunostaining and MFI and ROS levels in iPSC-RPE (n = 5) (D–F). (G–H) SA-β-gal staining of iPSC-RPE cells treated for 10 days (SO, 30 mM) and quantitative analysis of the percentage of SA-β-gal⁺ cells (n = 4). Scale bar: 20 µm. Data are expressed as mean ± SD. Statistical significance was determined by Mann–Whitney U tests (A, C, H) or unpaired two-tailed Student's t-tests (E, F). *P < 0.05; **P < 0.01; ns, not significant (P ≥ 0.05).

In summary, our data support a model in which iron overload is accompanied by a transient enhancement of glucose metabolic flux at early stages, potentially increasing mitochondrial burden and oxidative stress and thereby associating with senescence-like phenotypes in RPE cells. Attenuating excessive glucose-derived metabolic input partially restored mitochondrial homeostasis and reduced senescence-associated features, with pyruvate serving as a readout of altered metabolic flux. Collectively, these findings suggest that glucose metabolic reprogramming may modulate the susceptibility of RPE cells to iron-induced senescence and highlight metabolic flux regulation as a potential intervention point.

2-DG Ameliorates Iron-Induced RPE Senescence in Mice

To evaluate the therapeutic effect of reducing glucose metabolic flux in RPE senescence, we administered the glucose metabolism inhibitor 2-DG in the mouse model of iron overload. Consistent with our in vitro chronic exposure data, a 2-week FAC treatment altered metabolic flux in RPE–choroid complexes, as indicated by reduced pyruvate levels, which were further decreased by co-administration of 2-DG (Fig. 7A). Importantly, 2-DG alone caused no retinal toxicity, as indicated by unchanged ERG a- and b-wave amplitudes, and it significantly preserved visual function against FAC-induced impairment (Fig. 7B). 2-DG reduced the SA-β-gal–positive area in the RPE layer (Fig. 7C) and suppressed the mRNA expression of SASP genes, including IL6, CCL2, CXCL1, CXCL11, and SERPINE1, further supporting senescence attenuation (Fig. 7D). These results suggest that 2-DG modulates iron-associated RPE senescence in vivo and support a role for metabolic reprogramming, particularly glucose metabolic flux modulation, as a potential contributor to senescence regulation.

Figure 7.

2-DG alleviated iron-induced RPE cell senescence in a mouse model. Mice received weekly intravitreal injections of saline or FAC (100 µM, 1 µL) with or without intraperitoneal 2-DG (1000 mg/kg every 3 days) for 2 weeks. (A) Pyruvate levels in mouse RPE–choroid complex (n = 4). (B) ERG analysis of retinal function (n = 6). (C) SA-β-gal staining of mouse retinal sections and quantitative analysis of the percentage of SA-β-gal–positive area within the RPE layer (n = 4). (D) SASP gene expression in mouse RPE–choroid complex (n = 4). Scale bar: 20 µm. Data are expressed as mean ± SD. Statistical significance was determined by Kruskal–Wallis tests (A), one-way ANOVA test (B), or Mann–Whitney U tests (C, D). *P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant (P ≥ 0.05).

Discussion

Our findings support a conceptual framework in which iron overload is associated with altered glucose metabolic flux and elevated ROS levels, which correlate with disrupted mitochondrial homeostasis and may contribute to RPE cell senescence. Furthermore, modulation of glucose metabolic flux was sufficient to partially restore mitochondrial homeostasis and attenuate senescence-associated phenotypes in RPE cells.

Ferroptosis is an iron-dependent form of regulated cell death driven by lipid peroxidation and dysregulated iron metabolism. Ferroptosis has been associated with iron-induced lipid peroxidation and RPE degeneration in models of dry AMD.^33,34 In this study, our observations are more consistent with a senescence-associated stress response than overt ferroptotic cell death. Although canonical ferroptosis markers were not assessed, longitudinal PI staining revealed no significant cell death prior to the senescence endpoint, indicating that RPE cells remained largely viable under our experimental conditions. The FAC concentration used here (200 µM) was lower than doses frequently employed to trigger cell death (e.g., 500 µM),²⁸ consistent with our aim to examine cellular responses to iron stress. Notably, the iron overload models used in this study were moderate and chronic in nature, consisting of intravitreal FAC injections (1 µL of 0.1-mM weekly for 2 weeks) and intraperitoneal FAC injections (60 mg/kg every 3 days for 3 months), in contrast to previously reported acute or high-dose iron models that rapidly induce RPE and photoreceptor death.^35,36 Together, these observations support a threshold-dependent model in which chronic, low-intensity iron stress is associated with senescence-like responses, whereas prolonged exposure to iron may shift RPE cells toward ferroptosis.

RPE cell senescence has emerged as an important contributing factor in AMD pathogenesis, marked by lipofuscin buildup, mitochondrial decline, and chronic SASP-driven inflammation.^23,37,38 Due to their high metabolic demand, RPE cells are densely packed with mitochondria and represent major cellular ROS sources.^39,40 With aging, mitochondria not only produce more ROS but also become increasingly vulnerable to oxidative damage.^41,42 In AMD, donor RPE cells exhibit greater mitochondrial DNA (mtDNA) damage than age-matched controls, accompanied by reduced mitochondrial function and ATP output.^23,43 Dysfunctional mitophagy in RPE cells also contributes to AMD development by allowing damaged mitochondria to accumulate, which increases oxidative stress, chronic inflammation, and cellular senescence.^38,44,45 Notably, restoring mitochondrial activity can reverse AMD-like phenotypes in Elovl2-deficient RPE cells.⁴⁶ Previous work in ARPE-19 cells showed that ferritinophagy-mediated iron release triggers senescence via cyclic GMP–AMP synthase (cGAS)–stimulator of interferon genes (STING) activation,¹⁸ and in vivo iron overload induces AMD-like pathology. Another study linked mitochondrial ROS to DNA damage and senescence signaling in RPE.⁵ However, these studies did not address metabolic regulation. Compared to ARPE-19, which exhibits abnormal polarization, reduced phagocytic capacity, and impaired barrier function,^47–49 we employed iPSC-RPE cells, which retain polarized cytokine secretion and photoreceptor phagocytosis,^50,51 and iRPE cells, which offer improved transcriptomic stability.^28,29 These advanced models provided a robust platform for discovery, allowing us to better characterize the molecular associations underlying iron-induced RPE senescence.

Mitochondria are key organelles responsible for ROS production. During electrons transferring to oxygen via the electron transport chain (ETC), approximately 2% of molecular oxygen is diverted to ROS generation.⁵² During iron overload, increased labile iron can enhance radical generation via Fenton chemistry and impair iron–sulfur cluster assembly, leading to reduced ETC efficiency and increased electron leakage, which secondarily enhances mitochondrial ROS production.^53–55 Mitochondria act as the central hub of glucose metabolism, where pyruvate is transported into the mitochondrial matrix to fuel the TCA. Consistent with previous reports, enhanced glucose metabolic flux or increased pyruvate oxidation has been associated with elevated mitochondrial ROS and senescence phenotypes.^25–27 In line with these observations, our findings show that early inhibition of glucose metabolism with 2-DG attenuates ROS accumulation and partially restores mitochondrial homeostasis during iron treatment. Another line of evidence supporting the link between glucose metabolism and cellular senescence comes from studies showing that high-glucose conditions induce senescence in various cell types.²⁴ Mechanistically, glucose-induced senescence also involves N-glycosylation modification, which impairs protein function.⁵⁶ Collectively, glucose metabolic reprogramming represents a promising therapeutic strategy for inhibiting RPE cell senescence and treating AMD.

2-DG is a glucose analog that functions as a competitive inhibitor of glucose metabolism, primarily suppressing glycolysis through its inhibitory effect on hexokinase.⁵⁷ Beyond its metabolic effects, 2-DG exhibits anti-aging properties; for example, blockade of glycolysis by 2-DG delays melanocyte senescence while preserving cellular functions.⁵⁸ Although fatty acids constitute the primary energy source for RPE cells, glucose remains indispensable, as glucose deprivation in rat RPE cultures reduces cell viability by approximately 30%.⁵⁹ This underscores the critical role of glucose metabolism in maintaining RPE cell homeostasis. Interestingly, in addition to glycolysis inhibition, 2-DG treatment has been reported to enhance the pentose phosphate pathway (PPP), increasing NADPH production via glucose-6-phosphate dehydrogenase.⁶⁰ The resulting elevation in NADPH promotes the accumulation of reduced glutathione (GSH), as observed in NB4 cells,⁶¹ suggesting that 2-DG may suppress cellular senescence by bolstering cellular antioxidant capacity. Furthermore, 2-DG is also likely to attenuate RPE cell senescence by inhibiting N-glycosylation modifications.^56,62 These observations suggest that the protective effects of 2-DG are likely mediated through multiple, non-mutually exclusive mechanisms.

In conclusion, our study supports glucose metabolic reprogramming as a key contributing mechanism associated with iron-induced RPE senescence. We identified a transient increase in glucose metabolic activity as an early event correlated with mitochondrial dysfunction and ROS accumulation. Therapeutically, early modulation of this metabolic surge was associated with changes in senescence-related phenotypes, suggesting that glucose metabolic reprogramming contributes to iron-induced RPE dysfunction. These findings support further exploration of glucose metabolic modulation as a potential approach to maintain RPE homeostasis in AMD and other iron-related degenerative conditions.