Abstract
Fuchs endothelial corneal dystrophy (FECD), the leading indication for corneal transplantation in the U.S., causes loss of corneal endothelial cells (CECs) and corneal edema leading to vision loss. FECD pathogenesis is linked to impaired response to oxidative stress and environmental ultraviolet A (UVA) exposure. Although UVA is known to cause nonapoptotic oxidative cell death resulting from iron-mediated lipid peroxidation, ferroptosis has not been characterized in FECD. We investigated the roles of genetic background and UVA exposure in causing CEC degeneration in FECD. Using ungenotyped FECD patient surgical samples, we found increased levels of cytosolic ferrous iron (Fe2+) and lipid peroxidation in end-stage diseased tissues compared with healthy controls. Using primary and immortalized cell cultures modeling the TCF4 intronic trinucleotide repeat expansion genotype, we found altered gene and protein expression involved in ferroptosis compared to controls including elevated levels of Fe2+, basal lipid peroxidation, and the ferroptosis-specific marker transferrin receptor 1. Increased cytosolic Fe2+ levels were detected after physiologically relevant doses of UVA exposure, indicating a role for ferroptosis in FECD disease progression. Cultured cells were more prone to ferroptosis induced by RSL3 and UVA than controls, indicating ferroptosis susceptibility is increased by both FECD genetic background and UVA. Finally, cell death was preventable after RSL3 induced ferroptosis using solubilized ubiquinol, indicating a role for anti-ferroptosis therapies in FECD. This investigation demonstrates that genetic background and UVA exposure contribute to iron-mediated lipid peroxidation and cell death in FECD, and provides the basis for future investigations of ferroptosis-mediated disease progression in FECD.
Keywords: Corneal endothelium, Corneal transplant, Ferritin, Ferroptosis, Ferrous iron, Fuchs endothelial corneal dystrophy, Reactive oxygen species, Transferrin, Ubiquinol, Ultraviolet light
1. Introduction
Fuchs endothelial corneal dystrophy (FECD) is a complex age-related polygenic disease that affects roughly 6.1 million Americans [1] and represents the leading indication for corneal transplant surgery in the U.S [[2], [3], [4], [5], [6]]. The condition can be diagnosed early in early stages and well before it causes visual dysfunction, through the identification of degenerative extracellular matrix deposits (guttae) on the corneal endothelium that lines the inner cornea. Gradual progression of disease results in loss of corneal endothelial cells (CECs), which do not regenerate [7], and failure to maintain appropriate corneal hydration through active ion pumping to counter the passive leakage of aqueous humor [8]. Unfortunately, no available medical therapy can yet prevent disease progression, so advanced FECD requires transplantation of the endothelial cell layer to restore vision. On the molecular level, affected CECs have an increased steady-state level of reactive oxygen species (ROS), impaired antioxidant response to oxidative stress, and increased sensitivity to known exogenous stressors that drive disease progression including ultraviolet light (UV) [[9], [10], [11], [12]]. The cornea is particularly susceptible to damage by ultraviolet A light (UVA, 320–400 nm), which comprises the vast majority of incident solar radiation absorbed by CECs [9,10,13]. Unlike ultraviolet B (UVB) light that causes DNA damage directly, UVA light causes macromolecular damage indirectly via the production of ROS that results from irradiation [14]. DNA damage and apoptotic cell death in particular have been the focus of studies examining UV mediated cell death in FECD [9,[15], [16], [17]]. However, UVA is also known to result in lipid peroxidation and nonapoptotic oxidative cell death [18,19]. The roles of genetic background and UVA exposure in causing lipid membrane damage and endothelial cell degeneration in FECD have not been explored systematically. Specifically, the background effects of FECD mutation and exposure related effects of UVA irradiation on lipid peroxidation and CEC functioning in FECD patients or in vivo models have not been characterized.
A key characteristic of FECD is impaired endogenous response to oxidative stress. Vulnerability to oxidative damage in FECD has been well established, and factors that increase CEC susceptibility to lipid peroxidation have been described. Previous studies have reported decreased transcription of key antioxidant defenses in FECD including glutathione S-transferase, superoxide dismutase 2, aldehyde dehydrogenase 3A1, heme oxygenase 1, thioredoxin reductase 1, and several peroxiredoxins including Prdx 1, which protects against lipid peroxidation [[20], [21], [22], [23], [24], [25]]. Importantly, protein levels of nuclear factor erythroid 2–related factor 2 (NRF2) – the regulator of a wide-ranging metabolic response to oxidative stress, including the cystine/glutamate antiporter (system xc-) that imports cysteine for glutathione biosynthesis – are reduced in FECD [20,26,27]. Normal levels of glutathione (GSH) and normal functioning of glutathione peroxidase 4 (GPX4), which catalyzes the reduction of lipid peroxides in a GSH-dependent reaction, are important for protecting cells against nonapoptotic oxidative cell death via iron-dependent lipid peroxidation, or ferroptosis [[28], [29], [30]]. Of note, GPX4 levels are lower than controls in FECD surgical samples, indicating that FECD increases susceptibility to ferroptosis [31,32]. A possible role for ferroptosis in FECD has been postulated but not investigated to account for the increased susceptibility to oxidative damage and lipid peroxidation [4,33]. In addition to increased ROS-mediated lipid peroxidation and decreased GPX4 functioning, altered iron metabolism is required for ferroptosis to occur [34,35]. In health, ferric iron (Fe3+) is bound to transferrin, imported into the cell via transferrin receptor 1 (TFR1, also known as CD71) mediated endocytosis, and stored in ferritin [36,37]. A pool of labile and biologically reactive Fe2+ is available, but levels are carefully regulated in the cell [37,38]. In pathological exposures to oxidative stress, an excess of free Fe2+ reacts with membrane-bound lipids to cause ferroptosis [34]. Reactive Fe2+ at or near cell membranes can drive Fenton reactions, which cause the formation of toxic prooxidant radicals (unstable) and non-radical lipid hydroperoxide intermediaries (stable and detectable) and result in membrane-bound lipid peroxidation [[39], [40], [41], [42], [43]]. This is of particular interest in FECD given the increased susceptibility to UVA-induced damage in affected patients. UV exposure is known to result in iron accumulation and ferroptosis [[44], [45], [46]], and iron release from ferritin with UVA irradiation has been well reported in literature [45,47,48]. To date, no studies characterizing the role of iron, iron-lipid perturbations, or lipid peroxidation in FECD pathobiology have been reported.
Although multiple lines of evidence support the theoretical basis for ferroptosis to be a pathological component of cell death in FECD, ferroptosis has not been characterized or evaluated systematically in this disease. A detailed understanding of ferroptosis in FECD would facilitate the development of targeted pharmacological therapies directed at preventing oxidative damage and oxidative cell death. We hypothesized that both FECD genetic background and UVA increase CEC susceptibility to nonapoptotic oxidative cell death and lipid peroxidation through the accumulation of toxic intracellular concentrations of ferrous iron. To test this hypothesis, we studied ungenotyped FECD patient surgical samples and primary and immortalized cell culture models from FECD patients with pathological expansions of trinucleotide repeats in intron 3 of the TCF4 gene (the most common genotype associated with FECD [[49], [50], [51]]) and evaluated for increases in lipid peroxidation, cytosolic Fe2+, and susceptibility to ferroptosis attributable to genetic background and UVA exposure. We also evaluated for the expression of key genes and proteins associated with ferroptosis, including TFR1 (responsible for the influx of iron inside the cell), FPN1 (efflux transporter of intracellular iron), FSP1 (ferroptosis suppressor protein 1, inhibits ferroptosis by NAD(P)H-dependent reduction of ubiquinone to ubiquinol), FTH (ferritin heavy chain, converts Fe2+ to Fe3+ and stores iron to maintain homeostasis), FTL (ferritin light chain, iron reservoir and removes excess iron), and GPX4 (inhibits lipid peroxidation by converting hydroperoxides into lipid alcohols) and the capacity for molecules with anti-ferroptotic activity to prevent key cellular processes implicated in ferroptosis [36,[52], [53], [54], [55], [56]]. Results of this investigation provide a basis for future mechanistic investigations of ferroptotic cell death and the prevention of disease progression in FECD.
2. Materials and methods
2.1. Consent and tissue collection
All investigations at the University of Iowa and Mayo Clinic were carried out following the guidelines of the Declaration of Helsinki. All tissues were obtained with informed consent by patients or the donor's family or next of kin. Approval was not required for the deidentified donor corneal tissues in this investigation according to the Institutional Review Board (IRB) at the University of Iowa. For FECD samples, human corneal endothelial tissue was collected at the University of Iowa and Mayo Clinic at the time of endothelial keratoplasty from patients with advanced FECD that were enrolled in the Proteomic Analysis of Corneal Health Study (IRB 201603746) or Mayo Clinic Hereditary Eye Disease Study (IRB 06–007210), respectively. For control samples, human corneal endothelial tissue was obtained from human donor eyes provided by the Iowa Lions Eye Bank (ILEB, Coralville, IA) and Lions Gift of Sight Eye Bank (St. Paul, MN).
2.2. Materials
All required chemicals were procured from commercial sources and utilized without further purification process following the manufacturer's guidelines. Ubiquinol was procured from Sigma Aldrich (United States Pharmacopeia [USP] reference standard, USA). γ-cyclodextrin was purchased from CI America (Portland, OR). BODIPY™ 581/591C11 lipid fluorescent probe, Dihydroethidium (Hydroethidine), and SYTOX® Green nucleic acid stain dye were procured from ThermoFisher Scientific (Waltham, MA). Cytosolic ferrous iron (Fe2+) detection dye FerroOrange and mitochondrial ferrous iron (Fe2+) detection dye Mito-FerroGreen were purchased from Dojindo EU GmbH (Munich, Germany). LipidSpot™ 610 Lipid Droplet Stain was procured from Biotium, Inc., USA. All other solvents and reagents were analytical and cell culture grade.
2.3. Ferroptosis RNA-Seq data analysis
RNA-Seq datasets of corneal endothelial samples from 47 patients with FECD and 21 donor controls were obtained from publicly available datasets (SRA Accession Numbers: PRJNA445238, PRJNA524323, PRJNA597343). Pre-symptomatic controls with TCF4 trinucleotide repeats were excluded from the analysis [31,32,57]. Reads were aligned and gene counts were made using STAR [58], data quality was assessed using FastQC, gene expression was normalized, batch corrected, and determined using EdgeR [59]. Genes were excluded if there were less than 3 counts in 20 or more samples. Significant differences (FDR<0.05) were calculated using EdgeR. Gene Set Enrichment Analysis was conducted on known ferroptosis gene signatures [60]. For the FerrDB gene signature, all genes that were ferroptosis drivers, suppressors or markers were included, which comprised of 211 expressed genes. Heatmap and hierarchical clustering was conducted using the ComplexHeatMap, cluster, and dendextend packages in R.
2.4. Human FECD surgical tissue and human donor cornea tissue samples
Human corneal endothelial tissues were collected at the time of endothelial keratoplasty from patients with advanced FECD using standard surgical techniques at the University of Iowa. Immediately after removal of the endothelial cell-Descemet membrane tissue complex (EDM) from the eye, approximately 2/3rd of the excised tissue was immediately placed into a cryopreservation vial on dry ice and stored at −80 °C until further processing. The remaining 1/3rd of the excised tissue was sent in formalin for histopathological analysis to confirm the diagnosis of FECD. Human donor corneas were procured within 24 h of donor death and preserved in Optisol-GS storage media (Bausch & Lomb, Irvine, CA) at 4 °C. Experiments were conducted within 14 days of preservation. All donors were 50–75 years old and each cornea was inspected and evaluated following standard protocols and procedures of the Eye Bank Association of America (EBAA) and ILEB. EDM tissues were prepared by mounting donor corneas onto a 9.5 mm vacuum trephine (Barron Precision Instruments, LLC, MI, USA) and scoring the endothelium and Descemet membrane into the stroma. The EDM complex was visualized with 0.06 % trypan blue ophthalmic solution (VisionBlue, DORC International, Netherlands) and the tissue was carefully peeled away from the stroma and immediately stored at −80 °C until further processing.
2.5. Human TCF4 expanded repeat primary corneal endothelial cell culture
Primary HCECs were established as described previously [61]. Briefly, FECD and control corneal endothelium were placed individually in Opti-MEM (Gibco, Waltham MA) with 8 % fetal bovine serum (FBS; Gibco) overnight at 37 °C, dissociated with 0.02 % EDTA (Sigma, St Louis, MO) in phosphate buffered saline (PBS, Gibco) for 1h at 37 °C, and plated in a single well of a 6-well collagen IV-coated plate (Corning, Tewksbury, MA) containing Joyce's media [61]. Once cell proliferation reached approximately 70 %, cells were dislodged from the plate with 1X trypsin, isolated, and centrifuged at 500 g for 5 min. Pelleted cells were resuspended in Joyce's media and replated at a ratio of 1:3. HCECs were grown to confluence in Joyce's media (5–7 days) and then incubated in maturation media (human endothelial-SFM, 2 % FBS and 1X antibiotic/antimycotic) for 12 days prior to experimentation. Demographics of donor of primary cells with their TCF4 trinucleotide repeat expansion size are provided in Supplementary Table 1 [50].
2.6. Human corneal endothelial cell (HCEC-B4G12) and F35T cell culture
Healthy immortalized human corneal endothelial cells (HCEC-B4G12) were procured from Leibniz Institute DSMZ-German Collection of Microorganisms and Cell Culture GmbH, Germany, and FECD immortalized human corneal endothelial cells (F35T) were a generous gift of Dr. Albert Jun (Johns Hopkins University, Baltimore, MD). F35T cells were derived from a FECD patient expressing the TCF4 transcript with approximately 4500 CUG repeats in intron 3 (Supplementary Fig. 1). Both cell lines were cultured in Opti-MEM® I Reduced-Serum Media (ThermoFisher) supplemented with 5 ng/mL of human epidermal growth factor (hEGF, ThermoFisher), 20 ng/mL of nerve growth factor (NGF, Fisher Scientific), 200 mg/L of calcium chloride (Sigma-Aldrich), 50 μg/mL of gentamicin (ThermoFisher), 1 mL of Normocin™ (50 mg/mL, Invivogen), 0.08 % chondroitin sulfate (Sigma-Aldrich) and 8 % fetal bovine serum (HyClone Characterized, US origin). Growth media was optimized for F35T cells and used similarly for B4G12 cells to exclude media-based technical variability. Media was filtered with 0.22 μM PTFE filters prior to use. Cells were incubated at 37 °C with a continuous supply of 5 % CO2 and passaged at the confluence. Plastic surfaces of cell culture dishes were coated with commercial FNC Coating Mix (Athena Environmental Sciences, Inc., USA) to facilitate the adherence of the endothelial cells.
2.7. Real-time PCR analysis
Human FECD patient tissues samples (multiple tissues pooled together in each group) and donor cornea EDM complexes (multiple tissues pooled together in each group) were pelleted, and RNA was extracted and purified using RNeasy kit (Qiagen) according to manufacturer's instructions. Similarly, HCEC-B4G12 and F35T cells were pelleted, and RNA was extracted. 18 ng total RNA was reverse transcribed using High-Capacity cDNA Reverse Transcription kit (Applied Biosystems). qPCR was performed on the CFX Connect thermal cycler (Bio-Rad Laboratories, Inc) with 10 s melting, 30 s annealing/extension for 40 cycles. Melt curve analysis was performed at the end of each qPCR run to verify single product formation. ΔΔCt values were calculated between cell types normalized to 18S and statistical analysis was performed using Student's t-test. Primers used for qPCR are mentioned in Supplementary Table 2.
2.8. Western blot analysis
Human FECD patient tissues samples and donor cornea EDM complexes were lysed in RIPA buffer with protease inhibitor in pools of 3 tissues for 45 min on ice. Similarly, HCEC-B4G12 and F35T cells were pelleted, and protein lysates were extracted. 0.6 μg of total protein was loaded per capillary (DM-TP01, Protein Simple) and lysates were probed with antibodies directed at 4-HNE (STA-035, Cell Biolabs), GPX4 (MAB5457-SP, R&D Systems), NRF2 (PA5-14144, Invitrogen), FSP-1 (20886-1-AP, Proteintech), Ferroportin/SLC40A1 (PA5-G4232, Invitrogen) and TFR1 (MABS1982, Millipore) proteins. Protein expression was normalized to total protein (DM-TP01, Protein Simple) and compared between tissue and cell types.
2.9. Immunohistochemistry for TFR1 protein detection
Human surgical explant corneal endothelial tissue was fixed with 4 % paraformaldehyde buffered solution, pH 7.4, for 10 min at room temperature within 2–4 h after surgery. Donor corneal endothelial tissue representing healthy (non-FECD) control tissue was fixed following the same protocol within 2 weeks of tissue procurement. Samples were washed three times with 20 mM PBS, pH 7.4, then incubated with 0.1 % Triton X-100 in for 30 min for permeabilization. Blocking was performed for 1 h at room temperature in 20 mM PBS, pH 7.4, containing 2 % BSA, 5 % normal goat serum and 0.1 % Triton X-100. Incubation with primary mouse monoclonal anti-TFR1 antibody (clone 3B82A1, catalog# MABS1982, lot# 3519825, EMD Millipore, Burlington, MA, USA) diluted 1:250 with 0.2 % BSA, 1 % normal goat serum and 0.1 % Triton X-100 in 20 mM PBS, pH 7.4 was done for 16 h at 4 °C. Samples were then washed four times with 0.1 % Triton X-100 in 20 mM PBS, pH 7.4, and solution containing secondary AlexaFluor 568 goat anti-mouse antibody (1:1,000, ex/em = 579/603 nm, catalog# A11004, lot# 2447869, Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA) and 300 nM DAPI (Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA) in 0.1 % Triton X-100 was applied for 2 h at room temperature. Tissue was rinsed with 20 mM PBS, pH 7.4, and distilled water and mounted under Aquamount mounting medium (Thermo Fisher Scientific, Waltham, MA, USA). Images were collected by sequential confocal laser scanning microscopy with a Leica SP8 STED microscope (Leica Microsystems, Mannheim, Germany).
2.10. Cytosolic iron (Fe2+) detection
Human FECD patient tissues samples and donor cornea EDM complexes (cut into half and measured as two technical replicates) were digested with 0.2 % collagenase type II and 0.05 % hyaluronidase in reduced serum OptiMEM-I (Gibco-BRL, Grand Island, NY) media supplemented with 50 μg/mL gentamycin for 3h at 37 °C with frequent agitation on a tube rotator (model: 05-450-127, Fisher Scientific). The digestion was completed with 1 × 0.5 % trypsin-EDTA. Following digestion, cells were filtered with a 100 μM cell strainer to get primary cell suspension. Cells were washed once with Live Imaging Solution, centrifuged, and resuspended in 100 μL Live Imaging Solution. Cells were transferred to a 24-well plate and 300 μL of 1 μmol/L FerroOrange staining solution (ex/em = 543/580 nm) was added. Cells were incubated for 30 min (37 °C, 5 % CO2) and transferred to flow cytometer tubes. Fluorescence was measured using a flow cytometer (BD FACSCalibur™) and results were analyzed using FlowJo (BD Biosciences, USA).
2.11. Multidimensional protein identification technology (MudPIT) mass spectrometry
Aqueous humor samples from patients with FECD and patients without FECD were collected from patients during surgery. The filter-assisted sample preparation (FASP) method was used for preparing samples for digestion [62]. It was solubilized in a mix containing ionic detergent 5 % sodium deoxycholate (SDC), buffer 100 mM triethylammonium bicarbonate (TEAB) at pH 8.0, and 3 mM dithiothreitol (DTT). Samples were then sonicated, spun down, and finally the supernatant was transferred to a 30 kD MWCO filter (Millipore, MA, USA) and centrifuged for 30 min at 13,000 g. The filtrate was discarded, and the remaining sample was buffer exchanged with 1 % SDC and 100 mM TEAB at pH 8.0. Following buffer exchange, the sample was alkylated with 15 mM iodoacetamide and then digested overnight with trypsin at an enzyme to substrate ratio of 1:100 in a Thermo-Mixer at 1000 RPM at 37 °C. Peptides were collected by centrifugation and reversed-phase stop-and-go extraction (STAGE) tips were used for desalting approximately 20 μg of digested peptides [63]. Elution solvent was a mixture of 80 % acetonitrile and 5 % ammonium hydroxide. Desalted peptides were lyophilized in a SpeedVac (Thermo Fisher Scientific, MA, USA) for 1 h. Peptide samples were then analyzed by ultra-performance liquid chromatography coupled with tandem mass spectrometry (UPLC-MS/MS). The UPLC system was an Easy-nLC 1000 UHPLC system (Thermo Fisher Scientific) coupled with a quadrupole-Orbitrap mass spectrometer (Q-Exactive; Thermo Fisher Scientific). The column was 2 μM Thermo Easy Spray PepMap C18 column (Thermo Fisher Scientific) with 500 mm × 75 μM i.d. Two mobile phases were used, phase A was composed of 97.5 % MilliQ water, 2 % acetonitrile, and 0.5 % formic acid and phase B was composed of 99.5 % acetonitrile, and 0.5 % formic acid. The elution events were 0–210 min, 0–25 % B in A and 210–240 min, 25–80 % B in A. Nano-electrospray ionization (Thermo Easy Spray source; Thermo Fisher Scientific) was used at 50 °C with an electrospray voltage of 2.2 kV. Tandem mass spectra were acquired from the top 20 ions in the full scan in the range between 400 and 1200 m/z while dynamic exclusion was set to 15 s and singly-charged ions were excluded from the analysis. Isolation width was set to 1.6 Da and full MS and MS/MS resolution were set to 70,000 and 17,500, respectively. The normalized collision energy was 25 eV and automatic gain control was set to 2e5. Max fill MS and max fill MS/MS were set to 20 and 60 ms, respectively, and the underfill ratio was set to 0.1 %.
For identifying peptides, msconvert was used to convert RAW data files to mzML format [64] and then Peak Picker HiRes tool from the OpenMS framework was used to generate MGF files from mzML format [65]. Peptide identification searches required precursor mass tolerance of 10 parts per million, fragment mass tolerance of 0.02 Da, strict tryptic cleavage, up to 2 missed cleavages, variable modification of methionine oxidation, fixed modification of cysteine alkylation, and protein-level expectation value scores of 0.0001 or lower. Finally, MGF files were searched using up-to-date protein sequence libraries available from X!Tandem [66], UniProtKB, and OMSSA [67]. Identified protein intensities were normalized to total peptide hits per sample and scaled to logarithmic base 10. Partek Genomics Suite version 7.21.1119. (MO, USA) was used to determine statistically significant proteins (analysis of variance [ANOVA], p < 0.05). Pathway significance was ascertained by the number of proteins in the dataset in common with known proteins in a single pathway, as determined by the IPA database (Qiagen).
2.12. Basal level of ROS quantification
HCEC-B4G12 and F35T cells at 250,000 cells/well in 24-well plate were stained with 10 μM of dihydroethidium (DHE, ex/em = 518/606 nm) in 1 mL of Live Imaging Solution and incubated for 30 min (37 °C, 5 % CO2). DHE is a fluorescent probe that reacts with ROS and when gets oxidized emits red fluorescence. After incubation, cells were transferred to flow cytometer tubes. Fluorescence was measured using a flow cytometer (BD FACSCalibur) and data was analyzed using FlowJo.
2.13. Confocal microscopy
Confocal microscopy was conducted with multiple fluorescent probes, including the Dihydroethidium (Hydroethidine; DHE, ex/em = 518/606 nm) for fluorescent probe for imaging ROS, C11-Bodipy 581/591 for lipid peroxidation, FerroOrange for cytosolic labile iron (ex/em = 543/580 nm), Mito-FerroGreen for mitochondrial iron (ex/em = 505/535 nm), and LipidSpot™ 610 (ex/em = 592/638 nm) for lipid droplets. HCEC-B4G12 and F35T cells were seeded at 100,000 to 200,000 cells/chamber in 4-chambered coverglass slides (Thermo Scientific Nunc Lab-Tek) after coating the glass surface with FNC Coating Mix and incubating at 37 °C and 5 % CO2 18–72 h. After incubation, media was discarded, and cells were washed with Live Imaging Solution if necessary. For iron detection, washing at least twice was important to remove extracellular iron from the media. For lipid peroxidation imaging, cells were stained with 0.9 mL of 5 μM C11-Bodipy 581/591 fluorescent probe in Live Imaging Solution for 20 min and stain solution was discarded before adding 0.9 mL of fresh Live Imaging Solution. For cytosolic and mitochondrial iron imaging, after washing twice, 500 μL of 1 μmol/L FerroOrange and 5 μmol/L Mito-FerroGreen staining solutions were added to the cells and incubated for 30 min, separately. For ROS imaging, cells were stained with 0.9 mL of 10 μM Dihydroethidium fluorescent probe in Live Imaging Solution for 30 min. For lipid droplet imaging, cells were stained with 0.9 mL of 1X LipidSpot™ 610 fluorescent probe in cell culture media for 30 min. All incubations were conducted in a cell culture incubator at 37 °C and 5 % CO2. Immediately after incubation, live cells were imaged using a Leica SP8 confocal microscope with a 63X oil lens equipped with Leica Application Suite X (LAS X) operating software. For each filter, all images were taken at the same gain level and image capture settings for both HCEC-B4G12 and F35T cells. Images were analyzed using Image J.
2.14. Mitochondrial superoxide assay
HCEC-B4G12 and F35T cells (50,000/well and 25,000/well, respectively) were grown in 96 well plates until reaching confluency. Cells were exposed to MitoROS 580 dye (ab219943, Abcam) for 30 min, and mitochondrial superoxide was quantified using Infinite M Plex plate reader (Tecan Group, Ltd) with Ex/Em = 540/590 nm. As a positive assay control, additional B4G12 and F35T cells were treated with antimycin-A (AMA) for 30 min prior to and during the MitoROS 580 dye treatment, for a total of 60 min. AMA generates ROS by inhibiting complex III of the mitochondrial electron transport chain. Cells were fixed with 4 % formaldehyde buffered in 0.1 M PBS, pH 7.4 for 30 min, washed with PBS 3 times, and incubated with 300 nM DAPI for 30 min. DAPI-stained nuclei were counted in whole well images captured by Cytation 5 instrument (BioTek Instruments, Inc) using Gen5 Image Plus (version 3.10) software, and resulting cell counts were applied to obtain normalized MitoROS 580 RFU/cell values for each well. Average values were calculated per group and compared using Student's t-test.
2.15. Cytosolic iron (Fe2+) detection by flow cytometry using FerroOrange fluorescent probe
HCEC-B4G12 and F35T cells were cultured in T75 flasks until they reached confluency. Cells were trypsinized and washed twice with 1X DPBS buffer to remove residual trypsin and serum-containing media. The cell suspension was centrifuged at 230 g for 5 min and resuspended in Live Imaging Solution (Thermo Fisher). FerroOrange staining solution of 1 μmol/L was prepared following the manufacturer's guidelines. In 24 well plates, 150,000 cells in 100 μL Live Imaging Solution were added to each well and 300 μL of 1 μmol/L FerroOrange staining solution was added. For the control group, 300 μL of Live Imaging Solution was added to the cells. Then cells were incubated for 30 min (37 °C, 5 % CO2). After incubation, cells were transferred to flow cytometry tubes and fluorescence was measured by flow cytometry (BD FACSCalibur™). Data analysis was carried out using FlowJo software.
2.16. Mitochondrial iron (Fe2+) detection by flow cytometry using Mito-FerroGreen fluorescent probe
HCEC-B4G12 and F35T cells were cultured, trypsinized, washed, and resuspended in Live Imaging Solution following the same protocol as cytosolic iron detection. Mito-FerroGreen staining solution of 5 μmol/L was prepared following the manufacturer's guidelines. For the experimental group, 500 μL of 5 μmol/L Mito-FerroGreen staining solution was added to 150,000 cells in 100 μL of Live Imaging Solution in a 24-well plate. For the control group, 500 μL of Live Imaging Solution was added to the cells. After 30 min incubation (37 °C, 5 % CO2), fluorescence was measured using a flow cytometer (BD FACSCalibur™) and data was analyzed using FlowJo.
2.17. Cytosolic ferritin ELISA
HCEC-B4G12 and F35T cells were cultured in T75 flasks with 3 biological replicates. At confluence, the protein was extracted with 1X RIPA buffer (Sigma-Aldrich) supplied with EDTA-free Protease Inhibitor Cocktail (cOmplete™, Roche). Protein was stored at −80 °C until the ELISA assay was performed. Protein concentration in each biological replicate was measured by Pierce™ BCA Protein Assay Kit (Thermo Scientific™). Three technical replicates of 100 μg protein from each biological replicate diluted with diluent supplied by the manufacturer were added to the antibody-coated wells (Ferritin Human ELISA Kit, Invitrogen). All the steps were carried out following the manufacturer's protocol. Spectra Max plus 384 Microplate Spectrophotometer was used to measure the absorbance at 490 nm. The standard calibration curve of ferritin was prepared in duplicates.
2.18. Mitochondrial ferritin (MTFT) ELISA
Mitochondrial ferritin in HCEC-B4G12 and F35T cells were quantified using Immunotag™ Mitochondrial Ferritin ELISA kit (G-Biosciences, USA). Protein was extracted following the same protocol used for the cytosolic ferritin ELISA. Three technical replicates of 100 μg protein from each biological replicate were added to antibody-coated wells and all the experimental steps were conducted according to the manufacturer's guidelines without any modification. The standard calibration curve of FTMT was prepared in duplicate.
2.19. Lipid peroxidation basal level (C11-Bodipy 581/591) assay
HCEC-B4G12 and F35T cells were cultured in T75 flasks until they reached confluency. After detachment with trypsin, 400,000 cells were stained with 2 mL of 5 μM C11-Bodipy 581/591 fluorescent probe in Live Imaging Solution (Thermo Fisher) or left unstained (unstained control group), mixed by pipetting, and incubated in a cell culture incubator (37 °C and 5 % CO2) for 20 min. Oxidation of the polyunsaturated butadienyl portion of C11-Bodipy shifts the fluorescence emission peak from red (591 nm) to green (510 nm) to allow detection of lipid peroxidation in the membrane of cells. After staining, the cell suspension was transferred to 15 mL Falcon tubes. All 15 mL tubes were centrifuged at 230 g for 5 min, then stain solution was discarded, and the cells were resuspended in 0.6 mL of Live Imaging Solution. After cells were transferred to flow cytometer tubes, fluorescence was measured using a flow cytometer (BD FACSCalibur™) and results were analyzed using FlowJo.
2.20. Quantification of lipid droplets
HCEC-B4G12 and F35T cells at 250,000 cells/well in 24-well plate were stained with 1X of LipidSpot™ 610 in 1 mL of cell culture media and incubated for 30 min (37 °C, 5 % CO2). Following incubation, cells were transferred to flow cytometer tubes. Fluorescence was measured using a flow cytometer (BD FACSCalibur) and data was analyzed using FlowJo.
2.21. Time-lapse confocal imaging
HCEC-B4G12 and F35T cells were seeded at 100,000–150,000 cells/chamber in a 4-chambered glass bottom coverglass slide (Thermo Scientific Nunc Lab-Tek) after coating with FNC Coating Mix and incubated for 18 h (37 °C and 5 % CO2). Cells were treated with 1 μM of RSL3, 1X of LipidSpot™ 610 (ex/em = 592/638 nm) and 50 nM of SYTOX™ Green nucleic acid stain dye (ex/em = 504/523 nm) in 0.9 mL of cell culture media. In another experiment, cells were exposed with UVA at 1.5 J/cm2. Following UVA exposure, cells were treated with 1X of LipidSpot™ 610 and 50 nM of SYTOX™ Green nucleic acid stain dye in 0.9 mL of cell culture media. Time-lapse Z-stack imaging was performed for 17 h–48 h using a LSM 980 confocal microscope (Zeiss) with Airyscan 2 with 63X oil lens. During imaging, cells incubated in the chamber were maintained at 37 °C and 5 % CO2. Image analysis and video production were performed using Imaris 9.9 (Oxford Instruments).
2.22. Deferoxamine (DFO) iron chelation assay
HCEC-B4G12 and F35T cells were seeded in 96-well plates at 5000 cells/well. After 18 h of incubation at 37 °C and a continuous supply of 5 % CO2, cells were treated with 100 μM of DFO. After incubating for 24 h (37 °C and 5 % CO2), cells were washed once with 1X DPBS, and RSL3 at the doses of 1, 2, and 5 μM in DMSO were added to the respective cells while an equal amount of DMSO was added to the control cells, and incubated for 2, 4, 6 and 8 h (37 °C and 5 % CO2). After the RSL3 treatment of 2, 4, 6, and 8 h, at each time point, cells were gently washed once with 1X DPBS and MTS reagent of 20 μL (Cell Titer-96 Aqueous One Solution, Promega, USA) in 80 μL of cell culture media for 3 h (37 °C and 5 % CO2). After the incubation, the absorbance was measured at 490 nm using the Spectra Max plus 384 Microplate Spectrophotometer (Molecular Devices, Sunnyvale, CA) following the manufacturer's guidelines.
2.23. Preparation of solubilized ubiquinol
Solubilized ubiquinol was prepared by the kneading method where physical complexation was formed between ubiquinol and γ-cyclodextrin (γ-CD) following our previously published protocol [33]. Briefly, ubiquinol and γ-CD were mixed at the molar ratio of 1:10 and a hydro-alcoholic solution at 1:1 ratio was added to the mixture to form a semi-liquid paste. Mixing was continued for 1 h and then the paste was vacuum dried to yield the powdered complex.
2.24. Lipid peroxidation inhibition by solubilized ubiquinol
The lipid peroxidation assay was conducted using C11-Bodipy 581/591 fluorescent probe following the same seeding, washing, staining and measurement steps described in the basal level of lipid peroxidation assay, except cells were treated with solubilized ubiquinol then challenged with 1 μM of RSL3 or left untreated as controls. 18 h after seeding, media was removed and solubilized ubiquinol at concentrations of 1, 10, 50, and 100 μM diluted in cell culture media were added to the cells and incubated for 24 h (37 °C and 5 % CO2). After incubation, cells were washed twice with 1X DPBS, and 1 μM of RSL3 in DMSO was added except for the untreated and untreated-unstained control groups. Cells were then incubated for 8 h at 37 °C and 5 % CO2, and after the incubation, cells were stained with 2 μL of C11-Bodipy 581/591 fluorescent probe stock prepared in DMSO as described in the basal lipid peroxidation assay.
2.25. Ferroptosis assay using LDH
Lactate dehydrogenase (LDH) assay was performed using CyQUANT™ LDH Cytotoxicity Assay kit (Invitrogen, USA). LDH release in media was quantified as an endpoint of measuring ferroptosis. HCEC-B4G12 and F35T cells were seeded at 5000 cells/well in a 96-well plate and incubated for 18 h (37 °C and 5 % CO2). After incubation, media was removed, and solubilized ubiquinol dispersed in media was added at 1, 5, 10, 50, and 100 μM concentrations while only media was added to the control group. In this assay, one set of cells was used for measuring the spontaneous LDH activity, and another set of cells was used for measuring maximum LDH activity. Ferrostatin-1 (Sigma-Aldrich, USA), a known ferroptosis inhibitor, was dissolved in DMSO at 1 μM and used as an anti-ferroptotic positive control. After incubating for 24 h at 37 °C and 5 % CO2, cells were washed twice with 1X DPBS and treated with 1 μM of RSL3 in DMSO whereas control groups were only treated with an equal amount of DMSO in media. After 24 h of incubation, 10 μL of supplied 10X lysis buffer was added to the group designated for measuring the maximum LDH activity and incubated in the cell culture incubator for 45 min. Following 45 min incubation, 50 μL from each well was transferred to a new 96-well flat-bottom well plate. To test the assay performance, 50 μL of 1X LDH positive control was added to three wells which were used as the LDH positive control group. 50 μL of supplied reaction mixture was added and incubated for 30 min in the dark at room temperature (RT). 50 μL of stop solution was added to each well to stop the reaction. Absorbance was measured at 490 nm following the manufacturer's protocol. Percent cell viability was calculated using the formula mentioned below:
2.26. Ferroptosis assay using MTS
F35T cells were seeded at 2500 and 5000 cells/well in 96-well plate and incubated for 18 h (37 °C and 5 % CO2). After the incubation, media was removed, and solubilized ubiquinol, N-Acetylcysteine (NAC) and deferoxamine (DFO) dispersed in media were added separately at 1 and 10 μM concentrations while only media was added to the control group. Ferroptosis inhibitor ferrostatin-1 (Sigma-Aldrich, USA) in DMSO was used as a positive control. After adding treatments, cells were incubated for 24 h (37 °C and 5 % CO2). Cells were then washed twice with 1X DPBS to remove any residual treatments and 1 μM of RSL3 in DMSO was added to the cells except for the no RSL3 control group where only DMSO in media was added. After adding RSL3, cells were incubated for an additional 24 h (37 °C and 5 % CO2) and bright-field microscopic images of live cells were taken using the EVOS Cell Imaging System (ThermoFisher, USA). Cells were gently washed once with 1X DPBS and 20 μL of MTS reagent in 80 μL of media was added to the cells and incubated for 3 h (37 °C and 5 % CO2). After the incubation, the absorbance was measured at 490 nm following the manufacturer's guidelines.
2.27. Cytosolic iron (Fe2+) detection upon UVA irradiation
Human donor corneas stored in Optisol-GS® at 4 °C were procured as noted above. This experiment was performed in pairwise fashion, where the right eye was exposed to UVA irradiation and the left eye from the same donor was used as a control. Corneas were washed with sterile Hank's balanced salt solution (HBSS) and placed in 12-well plate in HBSS with endothelial side facing the UVA light source. Cells were exposed to UVA irradiation at the fluence of 5 J/cm2 using a Rayonet Photochemical Reactor (RPR-200, The Southern NE Ultraviolet Co., Brandford, CT). Initially, various UVA fluence levels of 5–25 J/cm2 with 5 J/cm2 increments were screened to find a safe UVA dose for cells by measuring cell viability after UVA irradiation using a MTS assay kit. In our UVA dose screening, we observed that doses above 5 J/cm2 are toxic to F35T cells; thus, a dose of 5 J/cm2 was chosen as a relatively safe UVA dose for measuring cell viability and sensitivity to UVA irradiation. Control corneas were treated identically except they were not subjected to UVA irradiation. Both UVA irradiated and non-irradiated control corneas were digested following the same procedures noted previously and primary cell suspensions were obtained. Cells were resuspended in 100 μL of Live Imaging Solution and transferred to a 24-well plate and 600 μL of 1 μmol/L FerroOrange staining solution was added to each well. Staining was performed for 30 min (37 °C and 5 % CO2). Fluorescence was measured using a flow cytometer (BD FACSCalibur) and data were analyzed using FloJo.
2.28. Cytosolic and mitochondrial iron (Fe2+) detection following UVA irradiation
HCEC-B4G12 and F35T cells were cultured and prepared following the same protocol as used for cytosolic iron detection. To quantify cytosolic iron, 400,000 cells in 300 μL of Live Imaging Solution were added to each well in a 24 well plate. The UVA irradiation instrument set-up was the same as noted previously for the irradiation of human corneas. Various doses of UVA irradiation of 1, 2, 4, and 8 J/cm2 fluences were delivered. UVA irradiated and non-irradiated control cells were stained with 1.2 mL of 1 μmol/L FerroOrange staining solution for 30 min in the incubator (37 °C, 5 % CO2). For the detection of mitochondrial iron, 20,000 cells in 200 μL of Live Imaging Solution were added to each well and irradiated with a UV dose of 1, 2, 4, and 8 J/cm2. UV irradiated and non-irradiated control cells were stained with 700 μL of 5 μmol/L Mito-FerroGreen staining solution for 30 min in an incubator (37 °C, 5 % CO2). The same volume of Live Imaging solution was added to the unstained control group, which was used to monitor background fluorescence signals. After incubation, fluorescence was measured using a flow cytometer (BD FACSCalibur) and data was analyzed using FlowJo.
2.29. Statistical analysis
All data were expressed as the mean ± standard error of the mean (SEM). Statistical analysis was performed using the two-tailed Student's t-test when the experimental group was only compared with the control group. One-way ANOVA followed by Tukey's post-hoc test was utilized when multiple groups were compared with each other. P-values of less than 0.05 were considered statistically significant. All experiments were carried out with at least 3 biological replicates and in technical triplicate. Both parametric and non-paramatric statistical analysis lead to the same conclusions, hence parametric statistical analysis was presented in this manuscript. Statistical analysis was carried out using GraphPad Prism.
3. Results
3.1. FECD surgical samples demonstrate key markers of ferroptosis
To gain insight into whether ferroptosis plays a role in FECD pathogenesis, we first performed an analysis of published RNA-Seq datasets of samples from 36 FECD and 8 control patients for the presence of genes known to be involved in ferroptosis [60]. Gene set enrichment analysis showed downregulation of a ferroptotic gene signature in FECD patients compared to control patients (FDR <0.01); however, a mixed model analysis that allows for both upregulated and downregulated genes was even more significantly enriched (FDR <0.001, Fig. 1A), implicating the involvement of ferroptosis in FECD pathogenesis.
Fig. 1.
FECD surgical tissues show key markers of ferroptosis. (A) Heatmap with hierarchical clustering for 211 genes from the FerrDB database that includes known driver, suppressor, and marker ferroptosis genes that were expressed in the RNA-Seq datasets. For each plot, “pearson” was used for the clustering distance and “complete” for the hierarchical clustering method. The location where the representative dataset was collected (Mayo, Russia, or UTSW) and mutation type (Control, no TCF4 repeats [No_Rep] or TCF repeats [TCF4_Rep]) are shown for each sample. (B) FSP1 mRNA and protein expression in FECD and control tissues. (C)FTH mRNA expression in control and FECD tissues. (D)GPX4 mRNA expression in FECD and control human tissues. (E) Ferroportin (FPN1) mRNA and protein expression in FECD and control tissues. (F) FTL mRNA expression in control and FECD tissues. (G) TFR1 mRNA and protein expression in control and FECD surgical tissues. (H) Representative immunohistochemistry images of TFR1 localization in non-FECD and FECD donor cornea tissues. (I) 4-HNE protein expression in human surgical samples from patients with FECD (n = 8). All data of mRNA and protein expression are shown as mean ± SEM for n = 12 (Control tissues, 8 pools of 3, each pool contained 3 tissues) and n = 24 (FECD tissues, 8 pools of 3, each pool contained 3 tissues). All the statistical comparisons were conducted using two-tailed, unpaired Student's t-test, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗∗p < 0.0001. Relative gene expression is normalized by β-actin. (J) Cytosolic Fe2+ in primary CECs isolated from healthy human donor corneas (n = 11, each cornea divided into 2 sections) and FECD surgical explants (n = 7). Data are shown as mean ± SEM; ∗∗p < 0.01, Student's t-test.
Next, we sought to validate these findings in our own corneal endothelial cell-Descemet membrane tissue samples resected from ungenotyped FECD patients (Supplementary Tables 3 and 4) by evaluating for expression of key genes and proteins associated with ferroptosis, including FPN1, FSP1, FTH, FTL, GPX4 and TFR1 [35,68,69]. Consistent with the published RNA-Seq FECD datasets, our FECD surgical samples showed that FSP1, FTH, and GPX4 gene expression was decreased by 0.6, 0.25, and 0.62-fold, respectively (Fig. 1B–D), and FPN1 gene expression was increased by 2.06-fold in (Fig. 1E). FTL and TFR1 gene expression was downregulated by 0.36 and 0.39-fold in our samples (Fig. 1F and G) but showed no difference in the published datasets. Consistent with the gene expression data, FPN1 showed a 2.83-fold increase in protein expression (p < 0.001, Fig. 1E), indicating a compensatory response by FECD to export higher labile iron in order to restore iron homeostasis. Interestingly, though FSP1 and TFR1 showed decreased gene expression, protein expression in our surgical samples was significantly upregulated by 1.38 and 1.50-fold, respectively (p < 0.05 and p < 0.01, respectively) (Fig. 1B and G). TFR1, which internalizes transferrin-iron complexes through endocytosis, serves as a useful and specific marker of ferroptosis because it contributes to higher iron intake and correlates with iron-mediated lipid peroxidation [36]. Notably, immunohistochemistry images show that TFR1 protein expression was higher and localized more to the surface of CECs in FECD donor tissues than non-FECD donor tissues (Fig. 1H).
In addition to key ferroptotic expression markers, we assessed whether key cellular processes implicated in ferroptosis, such as lipid peroxidation and accumulation of cytosolic ferrous iron [Fe2+], were increased in our FECD surgical samples. Accordingly, FECD surgical samples showed 1.3-fold greater accumulation of lipid peroxidation end products compared to control donor tissues (p < 0.01) (Fig. 1I, Supplementary Fig. 2) and 32 % more cytosolic Fe2+ content in isolated CECs (p < 0.01) (Fig. 1J–Supplementary Table 5).
Furthermore, we collected aqueous humor samples from 4 patients with and 4 patients without FECD (Supplementary Table 6) and conducted protein mass spectrometry analysis to look for further evidence of iron dysregulation. A total of 23,171 protein isoforms were identified, of which 3448 protein isoforms had significantly altered expression in FECD patient aqueous compared to non-FECD aqueous samples (Supplementary Tables 7–8) [70]. Ingenuity molecular pathway analysis of the statistically significant proteins (p < 0.05) between FECD and control aqueous humor identified ferroptosis as one of the top 15 most significantly represented pathways. Several isoforms of human transferrin were higher in FECD aqueous samples compared to controls (3.8 < Ratio <18.7; p < 0.015), consistent with previous findings that receptor-mediated endocytosis of the transferrin-Fe3+ complex is required for ferroptosis [71]. A complete list of differentially expressed proteins can be found in Supplementary Table 8.
3.2. TCF4 expanded repeat primary FECD cell cultures demonstrate key markers of ferroptosis
To assess the presence of key markers of ferroptosis in FECD patients with known disease secondary to abnormal expansion of trinucleotide repeats in TCF4, we utilized human primary CEC lines derived from FECD patients with genotype confirmed TCF4 expanded repeats. Primary cells from FECD patients with TCF4 expanded repeats had upregulated mRNA expression of the key ferroptosis biomarker TFR1 compared to non-FECD primary control cells (Fig. 2A), which correlates positively with increased TFR1 protein expression in FECD surgical samples. Both FSP1 and GPX4 mRNA, key indicators of lipid peroxidation and antioxidant dysregulation in ferroptosis, were upregulated in primary cells from FECD patients with TCF4 expanded repeats compared to controls (Fig. 2B and C). In addition to TFR1, both FTH and FTL mRNA were also upregulated in primary cells from FECD patients with TCF4 expanded repeats (Fig. 2D and E), indicating altered iron metabolism and disruption of iron homeostasis leading to toxic concentrations of labile intracellular Fe2+ which is also required for ferroptotic cell death [34,35]. Of note, contrasting responses in gene expression between primary cells and surgical tissues may reflect changes attributable to disease stage and/or model differences.
Fig. 2.
FECD primary and immortalized cell cultures showkey markersof ferroptosis. (A)TFR1 mRNA expression in non-FECD and FECD donor expanded TCF4 repeat expansion primary cells. (B) FSP1 mRNA expression in primary cells. (C) GPX4 mRNA expression in human expanded TCF4 repeat expansion primary cells. (D)FTH mRNA expression in non-FECD and FECD donor expanded TCF4 repeat expansion primary cells. (E)FTL mRNA expression in non-FECD and FECD donor primary cells. (F) Representation of median of the fluorescence of DHE showing significant difference in ROS between indicated cells. DHE (FL2 fluorescence) peak of F35T cells shifts to right when compared to B4G12 cells. (G) Representative confocal images showing fluorescence of DHE indicating ROS in the indicated cell lines. (H) Mitochondrial ROS quantified by MitoROS 580 dye in the indicated cells. Data are shown as mean ± SEM; n = 3; ∗∗∗∗p < 0.0001, one-way ANOVA, followed by Tukey's post-hoc test. AMA indicates antimycin-A. (I)GPX4 mRNA and protein expression in HCEC-B4G12 and F35T cells. (J) Basal level of lipid peroxidation in HCEC-B4G12 and F35T cells quantified by C11-BODIPY fluorescent probe using flow cytometry. Comparisons of median fluorescence of C11-BODIPY detected in HCEC-B4G12 and F35T cells (10,000 cells). Data are shown as mean ± SEM; n = 3; ∗∗∗∗p < 0.0001, Student's t-test. C11-BODIPY (FL1 fluorescence) peak of F35T cells shifts to right when compares to B4G12 cells. (K) Representative confocal images showing fluorescence of reduced and oxidized dye in the indicated cell lines. (L) 4-HNE protein expression in HCEC-B4G12 and F35T cells. All data of mRNA and protein expression are shown as mean ± SEM for n = 5–9 (B4G12), n = 5–7 (F35T) and n = 4 (both non-FECD and FECD donor primary cells). All the statistical comparisons were conducted using two-tailed, unpaired Student's t-test, ∗∗∗p < 0.001. Relative gene expression is normalized by β-actin.
3.3. Increased oxidative stress, lipid peroxidation, and iron overload in FECD cells
Due to the scarcity of surgical FECD tissue samples and limited capacity for protein analysis in primary TCF4 expanded repeat CEC cultures, we utilized an established TCF4 expanded repeat CEC line derived from an FECD patient (F35T) and a control CEC line (HCEC-B4G12) for additional experiments. As oxidative stress is highly implicated in FECD pathogenesis, we first recapitulated our and others work [3,10,12,33,72] by showing that TCF4 expanded repeat F35T cells had a higher basal ROS (p < 0.0001, Fig. 2F and G) and basal mitochondrial superoxide concentration (p < 0.0001, Fig. 2H) than control HCEC-B4G12 cells. Consistent with our findings from surgical samples, TCF4 expanded repeat F35T cells had 0.47-fold lower GPX4 expression (p < 0.001; Fig. 2I) than control HCEC-B4G12 cells and a commensurate 2.89-fold greater basal level of lipid peroxidation (p < 0.0001; Fig. 2J and K) and 1.46-fold greater accumulation of lipid peroxidation end products (p < 0.05) (Fig. 2L, Supplementary Fig. 3). GPX4 gene expression was upregulated by 2.27-fold in F35T cells when compared to B4G12 cells though protein was downregulated (Fig. 2I). In agreement with findings in TCF4 expanded repeat primary cells, TCF4 expanded repeat F35T cells showed upregulation of TFR1, FTH and FTL mRNA by 4.77, 2.37 and 2.64-fold, respectively (Fig. 3A and 3B-C), and FSP1 mRNA and FSP1 protein were upregulated by 2.51 and 1.61-fold in F35T cells, respectively (Fig. 3D), in comparison to control cells. As noted with primary cell cultures, the contrast between tissues and these cells may reflect changes attributable to disease stage and/or model differences. Additionally, higher mRNA expression and lower protein expression for a particular gene (e.g., GPX4) may indicate expression compensation attempts by the cell to produce more protein.
Fig. 3.
FECD immortalized cell culture showskey gene and protein expression changes related to ferroptosis. (A)TFR1 mRNA and protein expression in HCEC-B4G12 and F35T cells. (B)FTH mRNA expression in indicated cells. (C)FTL mRNA expression in indicated cells. (D) FSP1 mRNA and protein expression in the indicated cells. All data of mRNA and protein expression are shown as mean ± SEM for n = 5–9 (B4G12) and n = 5–7 (F35T). All the statistical comparisons were conducted using two-tailed, unpaired Student's t-test, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001. Relative gene expression is normalized by β-actin. (E) Representation of median of the fluorescence of FerroOrange showing significant difference in cytosolic Fe2+ between indicated cells. B4G12 and F35T cells were stained with FerroOrange fluorescent probe to quantify cytosolic Fe2+ by flow cytometry. A minimum 10,000 cells were quantified for measuring the fluorescence. (F) Cell population distribution after staining with FerroOrange showing cytosolic Fe2+ content. (G) Representation of confocal microscopy images of HCEC-B4G12 and F35T cells stained with FerroOrange fluorescent probe showing and comparing cytosolic Fe2+. (H) Representation of median of the fluorescence of Mito-FerroGreen showing significant difference in mitochondrial Fe2+ between indicated cells. HCEC-B4G12 and F35T cells were stained with Mito-FerroGreen fluorescent probe to quantify mitochondrial Fe2+ by flow cytometry. A minimum of 10,000 cells were quantified for measuring fluorescence. (I) Cell population distribution after staining with Mito-FerroGreen showing mitochondrial Fe2+ content. (J) Representation of confocal microscopy images of HCEC-B4G12 and F35T cells stained with Mito-FerroGreen fluorescent probe showing and comparing mitochondrial Fe2+. Data are represented as means ± SEM; n = 3; unpaired Student's t-test, ∗∗∗∗p < 0.0001. (K) Quantification of human ferritin (Ft) by ELISA in protein from HCEC-B4G12 and F35T cells. The human ferritin levels are presented as mean ± SEM; n = 9 (3 biological replicates × 3 technical replicates); ∗∗∗p < 0.001, Student's t-test. (L) Quantification of mitochondrial ferritin (FtMt) by ELISA in protein from HCEC-B4G12 and F35T cells. The mitochondrial ferritin levels are presented as mean ± SEM; n = 9 (3 biological replicates × 3 technical replicates); ∗∗∗∗p < 0.0001, Student's t-test. (M) Representation of cellular iron metabolism in ferroptosis process.
Additionally, TCF4 expanded repeat F35T cells showed a 3.3-fold greater cytosolic (p < 0.0001) and 10.9-fold greater mitochondrial (p < 0.0001) Fe2+ accumulation than control HCEC-B4G12 cells (Fig. 3E–J), which was associated with a 5.4-fold upregulation of TFR1 (p < 0.001, Fig. 3A) and a 20 % and 72 % decrease in total ferritin (p < 0.001, Fig. 3K) and mitochondrial ferritin (FtMt) protein (p < 0.0001, Fig. 3L), respectively. Of note, ferritin can store labile Fe2+ inside its multimeric shell in a nontoxic state (Fig. 3M) [73] and FtMt can safely redistribute toxic iron from the cytosol to the mitochondria and protect mitochondria from iron-mediated oxidative damage [[74], [75], [76]]. However, the presence of increased intracellular Fe2+ in the context of decreased ferritin and FtMt indicates intracellular overloading of toxic Fe2+. Moreover, iron overload due to both higher iron uptake associated with TFR1 overexpression and downregulation of ferritin protein via ferritinophagy, which further increases TFR1 expression, have both been reported to lead to accumulation of toxic Fe2+ and promote ferroptosis (Fig. 3M) [73]. With linkages established between TCF4 expanded repeat F35T cells and both TCF4 expanded repeat primary cells and ungenotyped surgical tissue samples from FECD patients, additional experiments were conducted using F35T cells in order to overcome limitations with sample availability.
3.4. Increased susceptibility to RSL3-induced ferroptosis in FECD cells
Increased ROS and lipid peroxidation in the setting of iron imbalance are the key hallmarks of ferroptosis; therefore, we evaluated whether further cellular perturbation with RSL3, an inhibitor of GPX4 peroxidase activity and well known inducer of ferroptosis [77], would promote greater than expected ferroptotic cell death in F35T cells compared to HCEC-B4G12 control cells. Following RSL3-induced ferroptosis, we observed a dramatic, early increase in the number of lipid droplets in both F35T and control cells (1–3 h), with F35T cells showing more droplets at baseline (p < 0.0001, Fig. 4A and B) and after treatment (Fig. 4C and D, Supplementary video 1A-B) [78]; this was followed by a rapid and dramatic decrease in visible droplets that occurred just prior to cell death (≥4 h) (Fig. 4D, Supplementary video 1A-B). Taken together, we and others [78] show that lipid droplets initially form to maintain cellular energy homeostasis and prevent lipotoxicity; however, when this compensatory mechanism becomes overwhelmed, cells undergo lipophagy to induce lipid release, lipid peroxidation and subsequent ferroptotic cell death. Consequently, F35T cells showed increased ferroptosis susceptibility, undergoing ferroptotic cell death approximately 1 h earlier than control cells as evidenced by both Sytox cell death assay and light microscopy, with the latter demonstrating typical ferroptosis morphological changes including cellular swelling, presence of lipid droplets at the perinuclear space at the time of death, and plasma membrane rupture (Fig. 4D, Supplementary video 1A-B and 2A-B). Additionally, F35T cells reached peak levels of cell death 1.5 h sooner than control cells, and displayed a characteristic ferroptosis wave-like pattern of propagating lipid damage and cell death (Supplementary figure 4) [79]. Of note, iron-chelation therapy with deferoxamine (DFO) did protect control cells from RSL3-induced ferroptosis between 4 and 8 h (p < 0.0001; Fig. 4E) but had minimal effect on F35T cells, providing further evidence of increased ferroptosis susceptibility in FECD cells.
Fig. 4.
FECD demonstrates higher lipid droplets and higher susceptibility to ferroptosis induced by RSL3. (A) Representation of median fluorescence of LipidSpot™ 610 showing significant difference in ROS between indicated cell lines. LipidSpot™ 610 (FL4 fluorescence) peak of F35T cells shifts to right when compared to B4G12 cells. (B) Representative confocal images showing fluorescence of LipidSpot™ 610 indicating lipid droplets in the indicated cell lines. In flow cytometer analysis, data are shown as mean ± SEM; n = 9 (3 biological replicates × 3 technical replicates); ∗∗∗∗p < 0.0001, Student's t-test. (C) Representative confocal images showing RSL3 induced lipid droplets in the indicated cells. (D) Time lapse confocal images of ferroptosis events induced by RSL3. (E) Deferoxamine (DFO) iron chelation assay where HCEC-B4G12 and F35T cells were treated with DFO for 24h and then challenged with RSL3 at 1, 2 and 5 μM for different durations of 2, 4, 6 and 8h. Cell viability was measured by MTS assay. Data are represented as mean ± SEM for three biological replicates; Student's t-test, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.
3.5. UVA exposure induces iron overload and ferroptosis in FECD cells
As UVA is a major component of sunlight that causes severe oxidative stress in exposed cells [80,81], increases labile cytosolic Fe2+ release from the core of cytosolic and mitochondrial ferritins [45,47,81], and is implicated in FECD pathogenesis and progression [9,[15], [16], [17]], we considered whether UVA may exert its toxic effects in susceptible FECD cells by inducing iron overload and promoting ferroptosis. To evaluate this, we first treated F35T cells with varying amounts of UVA radiation to determine a physiological dose of UVA irradiation (5 J/cm2) that minimally affected cell viability (Supplementary Fig. 5). Upon UVA irradiation, labile cytosolic Fe2+ was increased in a pair-wise comparison with control unexposed corneas (p = 0.06; Fig. 5A and B, Supplementary Table 9), confirming that even subthreshold levels of UVA irradiation can result in CEC iron release and intracellular accumulation.
Fig. 5.
FECD demonstrates higher iron overload and higher susceptibility to ferroptosis induced by UVA irradiation. (A) Cytosolic Fe2+ release in human donor corneas upon UVA irradiation at 5 J/cm2. Primary endothelial cells were isolated after UVA irradiation and stained with FerroOrange fluorescent probe. The experiment was conducted pairwise, where the left cornea was used as control and the right cornea was exposed to UVA. Data are shown as geometric mean FerroOrange (FL2 fluorescence) signals; n = 6 paired biological samples. (B) Comparison of geometric mean of FerroOrange (FL2 fluorescence) signals. Data are mean ± SEM, Paired two-tailed Student's t-test and n = 6 paired biological samples. (C) Representation of percent increase of labile cytosolic Fe2+ after UVA irradiation at the fluence of 1, 2, 4 and 8 J/cm2. Indicated cells were stained with FerroOrange fluorescent probe immediately post-UVA (Data are shown as mean ± SEM, n = 9; 3 biological replicates × 3 technical replicates; ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001, one-way ANOVA, followed by Tukey's post-hoc test). (D) Representation of percent increase of mitochondrial Fe2+ after UVA irradiation at the fluence of 1, 2, 4 and 8 J/cm2. Indicated cells were stained with Mito-FerroGreen fluorescent probe immediately post-UVA (Data are shown as mean ± SEM, n = 9; 3 biological replicates × 3 technical replicates; ∗∗∗∗p < 0.0001, one-way ANOVA, followed by Tukey's post-hoc test.). (E) Time lapse confocal images of ferroptosis events induced by UVA irradiation at 1.5 J/cm2. Representative confocal images showing disappearance of lipid droplets at the time of ferroptosis induced death in the indicated cells. SYTOX green indicates cell death. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)
Next, we exposed both F35T and HCEC-B4G12 control cells to various doses of UVA irradiation above and below the predetermined subthreshold level (1, 2, 4 and 8 J/cm2). Upon irradiation, F35T cells had a significantly (p < 0.0001) higher increase in cytosolic Fe2+ accumulation (20–60 %) when compared with control cells (0–8%) at all doses of UVA irradiation tested (Fig. 5C), indicating increased susceptibility of F35T cells to UVA-mediated toxicity. Similarly, F35T cells had a significantly higher percentage of mitochondrial Fe2+ accumulation than control cells at all doses of UVA irradiation (p < 0.0001, Fig. 5D). At higher dose levels of 4 and 8 J/cm2, we observed a decrease in mitochondrial Fe2+ levels in both cell lines compared to lower irradiances, suggestive of UVA-induced mitochondrial fragmentation [82,83] in which Fe2+ is released into the cytosol [17]. Following UVA exposure of 1.5 J/cm2, we observed strikingly similar morphological findings in F35T cells compared to control cells consistent with RSL3-induced ferroptosis, including a dramatic increase in lipid droplets at baseline and after exposure (Fig. 5E), a sharp decrease in lipid droplets prior to cell death (3–4 h), earlier onset of cell death (0.5 h before control cells as evidenced by both Sytox cell death assay and light microscopy), earlier peak levels of cell death (2.5 h sooner than control cells), and typical ferroptosis morphological changes including cellular swelling, presence of lipid droplets at the perinuclear space at the time of death, and plasma membrane rupture (Fig. 5E, Supplementary video 3A-B).
3.6. Ubiquinol prevents RSL3-induced ferroptosis in FECD and healthy cells
Next, in order to identify potential therapeutic options for FECD, we evaluated several pharmacological compounds known to protect against lipid peroxidation, iron overload, and ferroptosis, including solubilized ubiquinol [33], ferroptosis inhibitor ferrostatin-1, antioxidant N-acetyl cysteine (NAC), and iron-chelator deferoxamine (DFO). Strikingly, after 24 h of RSL3 treatment, solubilized ubiquinol outperformed all other tested compounds in reducing lipid peroxidation and preventing cell death in both cell lines (Fig. 6, Fig. 7), completely abrogating RSL3-induced ferroptotic cell death at low concentrations (1–5 μM) without any evidence of morphological cellular damage (Fig. 6A–C). Ferrostatin-1 provided a small but significant rescue at 1 μM in preventing cell death with complete rescue at 5–10 μM, whereas NAC and DFO had no effect at the tested conccentrations in preventing RSL3-induced ferroptosis at 24 h (Fig. 6D). Furthermore, F35T cells demonstrated significantly more cell death than control cells (p < 0.01, Fig. 5A and B) and reached peak levels of cell death sooner than control cells (Fig. 4D, Supplementary video 1A-B, 2A-B).
Fig. 6.
Solubilized ubiquinol gives protection against ferroptosis. (A) Cell viability assay was performed by quantifying LDH release after treatment with solubilized ubiquinol at different concentrations of 1, 5, 10, 50, and 100 μM. Following solubilized ubiquinol treatment, HCEC-B4G12 and F35T cells were challenged with 1 μM of RSL3 for 24 h. Data are shown as mean ± SEM; n = 3; ∗∗∗∗p < 0.0001, Student's t-test against RSL3 alone group. (B) Comparisons of cell viability of HCEC-B4G12 and F35T cells following the treatment of solubilized ubiquinol. Solubilized ubiquinol was comparatively less effective in F35T cells, indicating more ferroptosis compared to B4G12 cells. Data are shown as mean ± SEM; n = 3; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗∗p < 0.0001, Student's t-test. (C) Representative light microscopy images showing that solubilized ubiquinol at 1 μM dose inhibited RSL3 induced ferroptosis in HCEC-B4G12 and F35T cells. (D) Solubilized ubiquinol outperforms NAC, DFO and ferrostatin-1 in protecting F35T cells against RSL3 induced cell death in ferroptosis assay. Data are shown as mean ± SEM, n = 3; ∗∗∗∗p < 0.0001, Student's t-test.
Fig. 7.
Solubilized ubiquinol gives protection against lipid peroxidation. (A) Schematic showing RSL3 induced ferroptosis and role of solubilized ubiquinol to prevent lipid peroxidation and ferroptosis. (B) Solubilized ubiquinol prevents lipid peroxidation in HCEC-B4G12 and F35T cells induced by RSL3 in dose dependent manner detected by the peak shift of C11-BODPY fluorescent probe in flow cytometer analysis. (C) C11-BODIPY fluorescence signals detected by flow cytometer following treatments with solubilized ubiquinol at different concentrations of 1, 10, 50 and 100 μM. Values are mean ± SEM; n = 3; ∗∗∗∗p < 0.0001, Student's t-test against RSL3-untreated group. (D) Representation of cell population undergoing RSL3 induced lipid peroxidation following solubilized ubiquinol treatment at the indicated concentrations in HCEC-B4G12 and F35T cells. Solubilized ubiquinol decreased the number of cells undergoing lipid peroxidation in a concentration dependent manner. Values are mean ± SEM; n = 3.
Interestingly, solubilized ubiquinol treatment provided significant but moderate inhibition of lipid peroxidation at 1 μM in F35T cells with much greater peroxidation reduction in control cells, yet both cell types showed a robust rescue in cell viability (p < 0.0001; Fig. 7A–D). This suggests that F35T cells have the ability to survive with higher levels of lipid peroxidation than control cells and that a smaller reduction in lipid peroxidation levels can have a major impact in cell viability. Both cell lines returned to baseline lipid peroxidation levels following 5–10 μM of solubilized ubiquinol treatment (Fig. 7B–D). Higher doses of solubilized ubiquinol did not further lower lipid peroxidation levels below baseline but did provide a small increase in cell viability above 100 %, suggesting an improvement in cellular health and possibly increased proliferative capacity above basal levels in both cell lines (Fig. 7C). Following RSL3 treatment, increased lipid peroxidation was also correlated with a dramatic decrease in visible lipid droplets, suggesting increased susceptibility to RSL3-induced ferroptosis (Fig. 4, Fig. 7B-C).
Of note, ubiquinol is an endogenous antioxidant that is converted from coenzyme Q10 via FSP1, a GPX4 glutathione-independent inhibitor of lipid peroxidation and ferroptosis (Fig. 7A) [53,84]. Thus, we looked at FSP1 protein expression in our samples and found increased protein expression in both FECD surgical samples (p < 0.05; Fig. 1B) and F35T cells (p < 0.001; Fig. 3D) compared to controls, indicating endogenous activation of the ubiquinol pathway to help protect against cellular damage and death in FECD.
4. Discussion
In this study, we examined ungenotyped FECD patient surgical samples and utilized a variety of primary cell and immortalized cell culture models from FECD patients with pathological expansions of trinucleotide repeats in intron 3 of the TCF4 gene (the most common genotype that causes a late onset of disease) to characterize lipid peroxidation and key iron metabolites in FECD and determine whether ferroptosis plays a role in its pathogenesis [[49], [50], [51]]. Additionally, we compared changes in iron-lipid interactions that were attributable to genetic background and UVA irradiation against healthy controls, and evaluated the effects of ferroptosis suppressor molecules. We found evidence of elevated lipid peroxide and cytosolic Fe2+ levels in FECD samples and models, consistent with increased susceptibility to ferroptosis in FECD attributable to genetic background. We identified the involvement of Fe2+, a key mediator of reactive oxidation reactions, in gene-dependent susceptibility to UVA exposure and detected increases in both cytosolic and mitochondrial Fe2+ levels after irradiation. TFR1, which has been proposed as a specific ferroptosis marker because it is responsible for the influx of iron inside the cell [36], was found to be elevated in all samples and models and may be a useful biomarker of ferroptosis in this disease. Interestingly, of the inhibitory ferroptosis molecules evaluated (N-acetyl-l-cysteine as a GSH precursor, deferoxamine as a chelator, and ubiquinol as an antioxidant), only solubilized ubiquinol was found to prevent cell death after experimentally induced ferroptosis in this series. Taken altogether, this study presents the first lines of evidence that iron-mediated lipid peroxidation is linked cell death in FECD, and establishes a basis for future mechanistic investigations of ferroptosis prevention in FECD.
While the genetic background of FECD is complex and age of onset varies based on genotype, the clinically recognized hallmarks of this disease are highly conserved. CTG trinucleotide repeat expansions in intron 3 of the TCF4 gene are found in more than 70 % of patients with FECD; patients with this genotypic marker tend to develop “late-onset” disease and tend to be female (3- to 4:1 female to male ratio). Although the exact mechanism of disease remains incompletely understood, it is possible that expressed non-coding regions of mRNA are toxic to CECs. The remaining portion of patients have a mutation in one of at least 10 known genes with a disease-causing mutation that affect a variety of CEC functions, from abnormal collagen production to altered ion channel pump function. Remarkably, all forms of this disease are characterized by an increased susceptibility to oxidative stress, regardless of genotype, age of onset, or gender. Our analysis of surgical tissue explants collected at the time of endothelial keratoplasty for patients with FECD was performed on pooled samples from ungenotyped patients. This broad approach bears direct correlation with the FECD clinical phenotypes common to end-stage disease (loss of CECs, confluent guttae on the inner cornea, corneal edema and vision loss). It is noteworthy that we found evidence of lipid peroxidation, cytosolic Fe2+ accumulation, and GPX4 depletion occurring in tandem in surgical samples. Our findings, which are contextualized by the well-documented background of ROS accumulation, support the general presence of increased susceptibility to ferroptosis in patients with FECD. Although we gathered additional primary cell and immortalized cell culture evidence that the TCF4 trinucleotide repeat expansion increases ferroptosis susceptibility, further studies are needed to determine if these findings apply to other FECD genotypes equally.
Previously, we showed that ferroptosis can be induced experimentally in healthy immortalized human CECs using erastin, a direct inhibitor of system xc- and the GSH synthetic pathway. In this study, we found evidence that FECD is associated with constitutive increases in cytosolic Fe2+ and lipid peroxidation as well as decreased GPX4, specifically in primary and immortalized in vitro patient derived FECD cell culture models harboring CTG repeat expansions in TCF4. Multiple lines of evidence support the theoretical basis for ferroptosis to be a pathological component of cell death in FECD [4,12,26,85]. The TCF4 genotype models utilized in this study reproduce the ferroptosis pathway signatures found in both diseased tissue and direct experimental induction in healthy cells. Our findings are a strong indication that intronic repeat expansions likely render the FECD endothelium more susceptible to ferroptosis, and underscore the utility of these models – both primary and immortalized FECD cell lines, which returned similar data – in further studies of ferroptosis in FECD. They also suggest possible directions for investigating ferroptosis in other trinucleotide repeat expansions diseases, such as Huntington's.
In this study elevated levels of Fe2+, a reactive redox substrate under strict control and regulation, were detected in the cytosol of CECs from FECD patient samples and cell culture models. While the notion that iron overload plays a part in the pathobiology of Fuchs dystrophy is novel, it is well known that dysregulation of iron metabolism plays a role in several diseases. Typically, iron enters the cell from surrounding fluid after complexation with transferrin, a secretory product that binds Fe3+ (Fig. 8). The transferrin-Fe3+ complex enters via TFR1-mediated endocytosis, becomes acidified and reduced to Fe2+ in endosomes by STEAP3 metalloreductase, and gets released into the cytosol where it is converted to Fe3+ and incorporated into ferritin, a key iron storage protein and major regulator of intracellular iron [86,87]. Cells normally maintain low free cytosolic Fe2+ levels in a steady-state equilibrium with ferritin-bound iron because free Fe2+ is highly reactive and potentially toxic. Fe2+ can generate excessive ROS via Fenton reactions and increase the activity of lipoxygenases that are responsible for lipid peroxidation [81,86]. Thus, our finding of cytosolic Fe2+ accumulation in patient and cell culture models is noteworthy because it implicates ferroptosis as a mechanism of oxidative cell death in FECD. In ferroptosis, the sequence of Fe2+-mediated plasma membrane lipid peroxidation and subsequent cell death results in affected cells adopting a characteristic morphological appearance that features redistribution of lipids at the PM, occurring as lipid blebbing and bubble formation and leading toward cellular ballooning prior to death. Although we were unable to visualize ferroptosis cell morphology sequences in tissue explants owing to their terminal degeneration at the end stage of disease, we were able to visualize plasma membrane lipid degeneration and migration of pooled lipids toward the nuclear envelope in cultured FECD cells, confirming ferroptosis cell death sequences of affected cells.
Fig. 8.
Summary ofthemolecular mechanism of ferroptosis in FECD. Iron enters the cell in ferric form via TFR1-mediated endocytosis. Ferritin stores the excess iron in ferric form, which is nontoxic. The ferric form of iron gets converted to the ferrous form in endosomes. When labile, ferrous iron gets released into the cytosol, it causes lipid peroxidation via Fenton chemistry. UVA irradiation can cause iron release from ferritin which increases the labile iron pool in the cytosol, as well as increases iron-mediated lipid peroxidation, a process known as ferroptosis. GPX4 is the key regulator of ferroptosis, preventing occurrence through scavenging lipid peroxides and reactive oxygen species (ROS). In this study, RSL3 was used to block GPX4 to induce ferroptosis. Ubiquinol, the reduced and active form of coenzyme Q10, is a potent ferroptosis inhibitor that works by scavenging ROS and modulating iron metabolism. Ubiquinol is an essential participant in the FSP1-CoQ10-NAD(P)H pathway, an independent system working in parallel with GPX4 and glutathione to suppress lipid peroxidation and ferroptosis by supporting FSP1 function. Other molecules like DFO and artesunate can prevent ferroptosis by quenching labile toxic ferrous iron, however, are not solely as effective as ubiquinol in preventing ferroptosis.
We detected several points of disruption in iron regulation attributable to FECD genotype, including increased iron uptake signaling and storage disequilibrium. First, cultured FECD CECs with TCF4 repeat expansions show increased levels of ferroportin, the only membrane bound protein responsible for iron efflux [88]. FPN1 expression has been reported to be regulated by Hif1α, Hif2α and the Hif response element (HRE) under anemic conditions, which would cause increased expression to drive more iron into the blood from cells. More relevant to FECD, FPN1 expression is also regulated by NRF2 signaling in response to oxidative stress [89], where an increase in ferroportin expression indicates the cellular attempt to remove excess iron from the cytosol as we observed in these experiments. Of note, FPN1 transcription is also regulated by NRF2 in response to inflammation and in FECD there is evidence of inflammation [[90], [91], [92]]. In addition, increased amounts of transferrin, the secreted iron carrier protein necessary for uptake by cells, were detected in aqueous samples from patients with FECD at the time of transplant surgery. Increased transcription and expression of TFR1, the membrane bound receptor that endocytoses ferritin-Fe3+ complexes, was also detected in surgical tissue samples and all cell culture models. As well, decreased levels of cytosolic ferritin were found in cell culture models. The cell behavior observed with respect to iron uptake and storage is consistent with iron scavenging and decreased storage capacity, and these findings may direct future investigations into iron storage regulation and ferritin recycling via ferritinophagy. Interestingly, we also found a corresponding decrease in levels of mitochondrial ferritin, which like its cytosolic counterpart regulates mitochondrial iron and indicates the potential for Fe2+ overload in disease-affected CEC mitochondria. Mitochondrial impairment in FECD has been described extensively, with findings that include increased membrane depolarization, decreased ATP synthesis, decreased mtDNA, and increased mitophagy. In this study, we measured mitochondrial iron levels but did not measure mitochondrial membrane lipid peroxidation or other potential indicators of ferroptosis in FECD mitochondria. However, mitochondria may be particularly resistant to ferroptosis due to the presence of a third independent antioxidant defense enzyme, dihydroorotate dehydrogenase (DHODH), that is located on the outer face of the mitochondrial inner membrane and, like FSP1, directly supports the biosynthetic recycling of ubiquinol [93].
We found further evidence of the centrality of Fe2+ and importance of lipid peroxidation in FECD disease progression through UVA irradiation and ferroptosis suppression experiments. We first found that increased sensitivity to UV-induced oxidative damage in TCF4 FECD cell culture models was mediated by Fe2+ accumulation. After treating cultured cells with physiologic levels of UVA, we detected increased cytosolic levels of Fe2+ in FECD cells compared to controls treated with the same exposures; this response was dose dependent. UVA is a well-known risk factor for FECD disease progression, yet the mechanism by which UV causes oxidative damage in FECD has not been characterized previously. UVA is known to increase cytosolic Fe2+ by causing release from ferritin [44,45,94]. In skin, UVA causes lipid peroxidation in cell membranes via pathways involving Fe [81,95,96]; ferritin degradation within hours that results in immediate Fe2+ increases [94]; and compensatory increases in ferritin days after exposure [94,97]. In our study, UVA exposure caused cell death that was nearly identical to RSL3 induced ferroptosis, evidenced by the occurrence of early sharp increase of lipid droplets, dramatic decrease of lipid droplets at the onset of cell death, and presence of lipid droplets at the perinuclear space at the time of death. Our findings support current understandings of the role of ultraviolet as an accelerant of cell damage in FECD by ferroptosis, and demonstrate the involvement of Fe2+ in the mechanism of disease progression. We next found that iron chelation with DFO conferred some protection against ferroptosis induced by the direct inhibition of GPX4 using RSL3. When RSL3 inhibits GPX4, the labile iron pool causes lipid peroxidation and increases the accumulation of lipid ROS, resulting ferroptosis mediated cell death [68,77]. Therefore, a larger quantity of labile will increase the risk of ferroptosis when GPX4 is blocked by RSL3. F35T cells had significantly more cell death in comparison to HCEC-B4G12 cells at the same dose of RSL3 challenge and pre-treatment with DFO. DFO was less effective in preventing ferroptosis in F35T cells when compared with HCEC-B4G12 cells due to a comparatively higher labile iron pool. This experiment simultaneously confirmed the presence of increased basal intracellular iron concentrations and susceptibility to cell death in immortalized FECD cells, and provided confirmatory evidence that iron is involved in FECD-associated cell death.
We also found that ferroptosis suppression is possible in FECD cell cultures by co-treatment with ubiquinol. Ubiquinol, the reduced and active form of coenzyme Q10, is an essential participant in the FSP1-CoQ10-NAD(P)H pathway, an independent system working in parallel with GPX4 and glutathione to suppress lipid peroxidation and ferroptosis by supporting FSP1 function [36,53]. This experiment demonstrated the ability to prevent cell death after direct induction of ferroptosis using RSL3 in immortalized FECD cells using a previously validated formulation of ubiquinol complexed with γ-cyclodextrin that increases ubiquinol availability to CECs by increasing drug bioavailability in the aqueous phase. Interestingly, of the inhibitory ferroptosis molecules tested (NAC as a GSH precursor, DFO as a chelator, and ubiquinol as an antioxidant), solubilized ubiquinol clearly outperformed others in preventing cell death at the tested concentrations after experimentally induced ferroptosis. This resonates with our finding of increased FSP1 protein expression indicating endogenous activation of the ubiquinol pathway to help protect against cellular damage and death, and provides further support for ubiquinol targeted therapy to suppress ferroptosis and cell death in FECD. By demonstrating that specific anti-ferroptosis agents can mitigate the iron-mediated lipid peroxidation present in cells with this common FECD genetic background, we simultaneously confirm the importance of ferroptosis as a mechanism by which characteristic phenotypic damage occurs in CECs and establish the basis for investigating therapeutics to prevent ferroptosis-mediated damage and disease progression in FECD.
Limitations of this study include the use of primary expanded as well as immortalized FECD cell lines. Primary cell lines propagated from FECD patient samples are difficult to culture and provided limited cell numbers for our assays. Additionally, primary FECD cultured cells no longer truly convey the extent of disease in vitro; this is due to the growth medium and natural selection for cells that grow most efficiently to confluence, which are most likely the healthiest cells from the donor population. Also, immortalized cells are driven to propagate beyond their natural ability. We therefore did not anticipate that all measurements made in this investigation from in vitro samples to mirror those from in vivo samples. However, we used the cell lines successfully to better understand how the TCF4 repeat expansion genotype specifically affects ferroptosis susceptibility, and demonstrated that these cell culture models are useful tools for the study of ferroptosis in FECD.
5. Conclusion
Overall, this investigation presents the first lines of evidence that ferroptosis is a mechanism by which oxidative damage and CEC death occur in Fuchs endothelial corneal dystrophy. We present evidence that expanded trinucleotide repeats in intron 3 of TCF4 comprise a genetic background that results in increased cytosolic Fe2+, iron-mediated lipid peroxidation, and increased susceptibility to ferroptosis. We also present evidence that UVA exposure in cells with expanded repeats in TCF4 increases the risk of cell death via an iron-mediated mechanism, and highlight the importance of cytosolic Fe2+ accumulations as a plausible molecular mechanism for FECD disease progression. Experimental evidence demonstrating that both iron chelation and anti-ferroptosis antioxidant treatments prevent cell death in FECD cell cultures demonstrate the importance of these mechanisms. Although the evidence is strongest in this series for increased ferroptosis susceptibility in this TCF4 genotype, our findings apply to other FECD genotypes and further studies are needed to explore this possibility.
CRediT authorship contribution statement
Sanjib Saha: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Resources, Software, Validation, Visualization, Writing – original draft, Writing – review & editing. Jessica M. Skeie: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Project administration, Resources, Supervision, Visualization, Writing – original draft, Writing – review & editing. Gregory A. Schmidt: Investigation, Methodology, Resources, Writing – review & editing. Tim Eggleston: Formal analysis, Investigation, Methodology, Writing – review & editing. Hanna Shevalye: Formal analysis, Investigation, Methodology, Writing – review & editing. Christopher S. Sales: Investigation. Pornpoj Phruttiwanichakun: Formal analysis, Investigation, Writing – review & editing. Apurva Dusane: Investigation, Writing – review & editing. Matthew G. Field: Formal analysis, Investigation, Methodology, Visualization, Writing – original draft, Writing – review & editing. Tommy A. Rinkoski: Investigation, Writing – review & editing. Michael P. Fautsch: Investigation, Writing – review & editing. Keith H. Baratz: Investigation, Writing – review & editing. Madhuparna Roy: Investigation, Writing – review & editing. Albert S. Jun: Investigation, Writing – review & editing. Chandler Pendleton: Investigation, Writing – review & editing. Aliasger K. Salem: Conceptualization, Funding acquisition, Project administration, Resources, Supervision, Writing – review & editing. Mark A. Greiner: Conceptualization, Formal analysis, Funding acquisition, Investigation, Project administration, Resources, Supervision, Writing – original draft, Writing – review & editing.
Declaration of competing interest
Sanjib Saha, Declarations of interest: none.
Jessica M. Skeie, Declarations of interest: none.
Gregory A. Schmidt, Declarations of interest: none.
Tim Eggleston, Declarations of interest: none.
Hanna Shevalye, Declarations of interest: none.
Christopher S. Sales, Declarations of interest: none.
Pornpoj Phruttiwanichakun, Declarations of interest: none.
Apurva Dusane, Declarations of interest: none.
Matthew G. Field, Declarations of interest: none.
Tommy A. Rinkoski, Declarations of interest: none.
Michael P. Fautsch, Declarations of interest: none.
Keith H. Baratz, Declarations of interest: none.
Madhuparna Roy, Declarations of interest: none.
Albert S. Jun, Declarations of interest: none.
Chandler Pendleton, Declarations of interest: none.
Aliasger K. Salem, Declarations of interest: none.
Mark A. Greiner, Declarations of interest: none.
Acknowledgements
The authors cordially thank the Lyle and Sharon Bighley Endowed Chair of Pharmaceutical Sciences, Iowa Lions Eye Bank, the Beulah and Florence Usher Endowed Chair in Cornea/External Disease and Refractive Surgery, the M.D. Wagoner & M.A. Greiner Cornea Excellence Fund, Mr. and Mrs. Robert and Joell Brightfelt, Mr. and Mrs. Lloyd and Betty Schermer, the UIHC Cornea Research Fund for financial support, and the cornea patients, donors, and donor families that made this research possible. MAG, AKS, JMS and SS are supported by the NIH grant R21EY034198.
List of abbreviations
- DFO
Deferoxamine
- DHE
Dihydroethidium
- EDM
Endothelial cell-Descemet membrane
- FASP
Filter assisted sample preparation
- FECD
Fuchs endothelial corneal dystrophy
- ILEB
Iowa Lions Eye Bank
- IRB
Institutional Review Board
- LDH
Lactate dehydrogenase
- MudPIT
Multidimensional protein identification technology
- NAC
N-Acetylcysteine
- NGF
Nerve growth factor
- ROS
Reactive oxygen species
- SEM
Standard error of the mean
- TFRA
transferrin receptor 1
- USP:
United States Pharmacopeia
- UVA
Ultraviolet A
Footnotes
Supplementary data to this article can be found online at https://doi.org/10.1016/j.redox.2024.103348.
Contributor Information
Aliasger K. Salem, Email: aliasger-salem@uiowa.edu.
Mark A. Greiner, Email: mark-greiner@uiowa.edu.
Appendix A. Supplementary data
The following are the Supplementary data to this article:
References
- 1.Ołdak M., et al. Fuchs endothelial corneal dystrophy: strong association with rs613872 not paralleled by changes in corneal endothelial TCF4 mRNA level. BioMed Res. Int. 2015;2015 doi: 10.1155/2015/640234. (in eng) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Uchida T., Sakai O., Imai H., Ueta T. Role of glutathione peroxidase 4 in corneal endothelial cells. Curr. Eye Res. 2017/03/04 2017;42(3):380–385. doi: 10.1080/02713683.2016.1196707. [DOI] [PubMed] [Google Scholar]
- 3.Lovatt M., Adnan K., Peh G.S.L., Mehta J.S. Regulation of oxidative stress in corneal endothelial cells by Prdx6. Antioxidants. Dec 4 2018;7(12) doi: 10.3390/antiox7120180. (in eng) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Lovatt M., Adnan K., Kocaba V., Dirisamer M., Peh G.S.L., Mehta J.S. Peroxiredoxin-1 regulates lipid peroxidation in corneal endothelial cells. Redox Biol. 2020;30 doi: 10.1016/j.redox.2019.101417. (in eng) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Eye Bank Association of America . vol. 2021. 2020. (2019 Eye Banking Statistical Report). Washington, DC. [Google Scholar]
- 6.Wagoner M.D., et al. Feeder-free differentiation of cells exhibiting characteristics of corneal endothelium from human induced pluripotent stem cells. Biol Open. May 8 2018;7(5) doi: 10.1242/bio.032102. (in eng) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Joyce N.C. Proliferative capacity of corneal endothelial cells. Exp. Eye Res. 2012;95(1):16–23. doi: 10.1016/j.exer.2011.08.014. (in eng) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Jalimarada S.S., Ogando D.G., Bonanno J.A. Loss of ion transporters and increased unfolded protein response in Fuchs' dystrophy. Mol. Vis. 2014;20:1668–1679. https://pubmed.ncbi.nlm.nih.gov/25548511 https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4265779/ (in eng) [Online]. Available: [PMC free article] [PubMed] [Google Scholar]
- 9.Liu C., et al. Ultraviolet A light induces DNA damage and estrogen-DNA adducts in Fuchs endothelial corneal dystrophy causing females to be more affected. Proc. Natl. Acad. Sci. U. S. A. Jan 7 2020;117(1):573–583. doi: 10.1073/pnas.1912546116. (in eng) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Liu C., Vojnovic D., Kochevar I.E., Jurkunas U.V. UV-A irradiation activates nrf2-regulated antioxidant defense and induces p53/caspase3-dependent apoptosis in corneal endothelial cells. Invest. Ophthalmol. Vis. Sci. Apr 1 2016;57(4):2319–2327. doi: 10.1167/iovs.16-19097. (in eng) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.White T.L., Deshpande N., Kumar V., Gauthier A.G., Jurkunas U.V. Cell cycle re-entry and arrest in G2/M phase induces senescence and fibrosis in Fuchs Endothelial Corneal Dystrophy. Free Radic. Biol. Med. Feb 20 2021;164:34–43. doi: 10.1016/j.freeradbiomed.2020.12.445. (in eng) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Jurkunas U.V., Bitar M.S., Funaki T., Azizi B. Evidence of oxidative stress in the pathogenesis of fuchs endothelial corneal dystrophy. Am. J. Pathol. 2010;177(5):2278–2289. doi: 10.2353/ajpath.2010.100279. (in eng) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Zinflou C., Rochette P.J. Ultraviolet A-induced oxidation in cornea: characterization of the early oxidation-related events. Free Radic. Biol. Med. Jul 2017;108:118–128. doi: 10.1016/j.freeradbiomed.2017.03.022. (in eng) [DOI] [PubMed] [Google Scholar]
- 14.Delic N.C., Lyons J.G., Di Girolamo N., Halliday G.M. Damaging effects of ultraviolet radiation on the cornea. Photochem. Photobiol. Jul 2017;93(4):920–929. doi: 10.1111/php.12686. (in eng) [DOI] [PubMed] [Google Scholar]
- 15.Czarny P., et al. DNA damage and repair in Fuchs endothelial corneal dystrophy. Mol. Biol. Rep. Apr 2013;40(4):2977–2983. doi: 10.1007/s11033-012-2369-2. (in eng) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Halilovic A., et al. Menadione-induced DNA damage leads to mitochondrial dysfunction and fragmentation during rosette formation in fuchs endothelial corneal dystrophy. Antioxidants Redox Signal. Jun 20 2016;24(18):1072–1083. doi: 10.1089/ars.2015.6532. (in eng) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Kumar V., Jurkunas U.V. Mitochondrial dysfunction and mitophagy in fuchs endothelial corneal dystrophy. Cells. Jul 26 2021;10(8) doi: 10.3390/cells10081888. (in eng) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Moysan A., Marquis I., Gaboriau F., Santus R., Dubertret L., Morlière P. Ultraviolet A-induced lipid peroxidation and antioxidant defense systems in cultured human skin fibroblasts. J. Invest. Dermatol. May 1993;100(5):692–698. doi: 10.1111/1523-1747.ep12472352. (in eng) [DOI] [PubMed] [Google Scholar]
- 19.Vats K., et al. Keratinocyte death by ferroptosis initiates skin inflammation after UVB exposure. Redox Biol. 2021/11/01/2021;47 doi: 10.1016/j.redox.2021.102143. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Jurkunas U.V., Bitar M.S., Funaki T., Azizi B. Evidence of oxidative stress in the pathogenesis of fuchs endothelial corneal dystrophy. Am. J. Pathol. Nov 2010;177(5):2278–2289. doi: 10.2353/ajpath.2010.100279. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Jurkunas U.V., et al. Decreased expression of peroxiredoxins in Fuchs' endothelial dystrophy. Investigative ophthalmology & visual science. 2008;49(7):2956–2963. doi: 10.1167/iovs.07-1529. (in eng) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Gottsch J.D., et al. Serial analysis of gene expression in the corneal endothelium of Fuchs' dystrophy. Invest. Ophthalmol. Vis. Sci. Feb 2003;44(2):594–599. doi: 10.1167/iovs.02-0300. (in eng) [DOI] [PubMed] [Google Scholar]
- 23.Buddi R., Lin B., Atilano S.R., Zorapapel N.C., Kenney M.C., Brown D.J. Evidence of oxidative stress in human corneal diseases. J. Histochem. Cytochem. : official journal of the Histochemistry Society. Mar 2002;50(3):341–351. doi: 10.1177/002215540205000306. (in eng) [DOI] [PubMed] [Google Scholar]
- 24.Lovatt M., Kocaba V., Hui Neo D.J., Soh Y.Q., Mehta J.S. Nrf2: a unifying transcription factor in the pathogenesis of Fuchs' endothelial corneal dystrophy. Redox Biol. 2020/10/01/2020;37 doi: 10.1016/j.redox.2020.101763. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Kuang F., Liu J., Tang D., Kang R. Oxidative damage and antioxidant defense in ferroptosis. Front. Cell Dev. Biol. 2020;8 doi: 10.3389/fcell.2020.586578. (in eng) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Liu C., Chen Y., Kochevar I.E., Jurkunas U.V. Decreased DJ-1 leads to impaired Nrf2-regulated antioxidant defense and increased UV-A-induced apoptosis in corneal endothelial cells. Invest. Ophthalmol. Vis. Sci. Jul 31 2014;55(9):5551–5560. doi: 10.1167/iovs.14-14580. (in eng) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Dodson M., Castro-Portuguez R., Zhang D.D. NRF2 plays a critical role in mitigating lipid peroxidation and ferroptosis. Redox Biol. May 2019;23 doi: 10.1016/j.redox.2019.101107. (in eng) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Yang Wan S., et al. Regulation of ferroptotic cancer cell death by GPX4. Cell. 2014/01/16/2014;156(1):317–331. doi: 10.1016/j.cell.2013.12.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Forcina G.C., Dixon S.J. GPX4 at the crossroads of lipid homeostasis and ferroptosis. Proteomics. Sep 2019;19(18) doi: 10.1002/pmic.201800311. (in eng) [DOI] [PubMed] [Google Scholar]
- 30.Yang W.S., Stockwell B.R. Synthetic lethal screening identifies compounds activating iron-dependent, nonapoptotic cell death in oncogenic-RAS-harboring cancer cells. Chem. Biol. 2008/03/21/2008;15(3):234–245. doi: 10.1016/j.chembiol.2008.02.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Nikitina A.S., et al. Dataset on transcriptome profiling of corneal endothelium from patients with Fuchs endothelial corneal dystrophy. Data Brief. 2019;25 doi: 10.1016/j.dib.2019.104047. (in eng) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Wieben E.D., et al. Gene expression in the corneal endothelium of Fuchs endothelial corneal dystrophy patients with and without expansion of a trinucleotide repeat in TCF4. PLoS One. 2018;13(7) doi: 10.1371/journal.pone.0200005. (in eng) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Naguib Y.W., et al. Solubilized ubiquinol for preserving corneal function. Biomaterials. 2021/08/01/2021;275 doi: 10.1016/j.biomaterials.2021.120842. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Dixon S.J., et al. Ferroptosis: an iron-dependent form of nonapoptotic cell death. Cell. May 25 2012;149(5):1060–1072. doi: 10.1016/j.cell.2012.03.042. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Li J., et al. Ferroptosis: past, present and future. Cell Death Dis. 2020/02/03 2020;11(2):88. doi: 10.1038/s41419-020-2298-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Feng H., et al. Transferrin receptor is a specific ferroptosis marker. Cell Rep. 2020/03/10/2020;30(10):3411–3423. doi: 10.1016/j.celrep.2020.02.049. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Bogdan A.R., Miyazawa M., Hashimoto K., Tsuji Y. Regulators of iron homeostasis: new players in metabolism, cell death, and disease. Trends Biochem. Sci. 2016/03/01/2016;41(3):274–286. doi: 10.1016/j.tibs.2015.11.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Venkataramani V. Iron homeostasis and metabolism: two sides of a coin. Adv. Exp. Med. Biol. 2021;1301:25–40. doi: 10.1007/978-3-030-62026-4_3. (in eng) [DOI] [PubMed] [Google Scholar]
- 39.Girotti A.W. Lipid hydroperoxide generation, turnover, and effector action in biological systems. J. Lipid Res. Aug 1998;39(8):1529–1542. https://www.ncbi.nlm.nih.gov/pubmed/9717713 [Online]. Available: [PubMed] [Google Scholar]
- 40.Girotti A.W., Kriska T. Role of lipid hydroperoxides in photo-oxidative stress signaling. Antioxid Redox Signal. Apr 2004;6(2):301–310. doi: 10.1089/152308604322899369. [DOI] [PubMed] [Google Scholar]
- 41.Aust S.D., Morehouse L.A., Thomas C.E. Role of metals in oxygen radical reactions. J. Free Radic. Biol. Med. 1985;1(1):3–25. doi: 10.1016/0748-5514(85)90025-x. [DOI] [PubMed] [Google Scholar]
- 42.Clemente S.M., Martinez-Costa O.H., Monsalve M., Samhan-Arias A.K. Targeting lipid peroxidation for cancer treatment. Molecules. Nov 5 2020;25(21) doi: 10.3390/molecules25215144. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Halliwell B., Gutteridge J.M. Biologically relevant metal ion-dependent hydroxyl radical generation. An update. FEBS Lett. Jul 27 1992;307(1):108–112. doi: 10.1016/0014-5793(92)80911-y. [DOI] [PubMed] [Google Scholar]
- 44.Aubailly M., Santus R., Salmon S. Ferrous ion release from ferritin by ultraviolet-A radiations. Photochem. Photobiol. Nov 1991;54(5):769–773. doi: 10.1111/j.1751-1097.1991.tb02088.x. [DOI] [PubMed] [Google Scholar]
- 45.Wolszczak M., Gajda J. Iron release from ferritin induced by light and ionizing radiation. Res. Chem. Intermed. 2010/09/01 2010;36(5):549–563. doi: 10.1007/s11164-010-0155-0. [DOI] [Google Scholar]
- 46.Zhu J., et al. Upconverting nanocarriers enable triggered microtubule inhibition and concurrent ferroptosis induction for selective treatment of triple-negative breast cancer. Nano Lett. Sep 9 2020;20(9):6235–6245. doi: 10.1021/acs.nanolett.0c00502. [DOI] [PubMed] [Google Scholar]
- 47.Smith M.J., Fowler M., Naftalin R.J., Siow R.C.M. UVA irradiation increases ferrous iron release from human skin fibroblast and endothelial cell ferritin: consequences for cell senescence and aging. Free Radic. Biol. Med. 2020/08/01/2020;155:49–57. doi: 10.1016/j.freeradbiomed.2020.04.024. [DOI] [PubMed] [Google Scholar]
- 48.Vile G.F., Tyrrell R.M. Oxidative stress resulting from ultraviolet A irradiation of human skin fibroblasts leads to a heme oxygenase-dependent increase in ferritin. J. Biol. Chem. 1993/07/15/1993;268(20):14678–14681. doi: 10.1016/S0021-9258(18)82386-9. [DOI] [PubMed] [Google Scholar]
- 49.Afshari N.A., et al. Genome-wide association study identifies three novel loci in Fuchs endothelial corneal dystrophy. Nat. Commun. 2017/03/30 2017;8(1) doi: 10.1038/ncomms14898. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50.Wieben E.D., et al. Comparison of TCF4 repeat expansion length in corneal endothelium and leukocytes of patients with Fuchs endothelial corneal dystrophy. PLoS One. 2021;16(12) doi: 10.1371/journal.pone.0260837. (in eng) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51.Fautsch M.P., et al. TCF4-mediated Fuchs endothelial corneal dystrophy: insights into a common trinucleotide repeat-associated disease. Prog. Retin. Eye Res. Mar 2021;81 doi: 10.1016/j.preteyeres.2020.100883. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52.Chen X., Yu C., Kang R., Tang D. Iron metabolism in ferroptosis. Review. 2020-October-07 2020;8 doi: 10.3389/fcell.2020.590226. (in English) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 53.Doll S., et al. FSP1 is a glutathione-independent ferroptosis suppressor. Nature. 2019/11/01 2019;575(7784):693–698. doi: 10.1038/s41586-019-1707-0. [DOI] [PubMed] [Google Scholar]
- 54.Hu W., et al. FTH promotes the proliferation and renders the HCC cells specifically resist to ferroptosis by maintaining iron homeostasis. Cancer Cell Int. 2021/12/29 2021;21(1):709. doi: 10.1186/s12935-021-02420-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 55.Ke S., Wang C., Su Z., Lin S., Wu G. Integrated analysis reveals critical ferroptosis regulators and FTL contribute to cancer progression in hepatocellular carcinoma. Original Research. 2022-May-16 2022;13 doi: 10.3389/fgene.2022.897683. (in English) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 56.Seibt T.M., Proneth B., Conrad M. Role of GPX4 in ferroptosis and its pharmacological implication. Free Radic. Biol. Med. 2019/03/01/2019;133:144–152. doi: 10.1016/j.freeradbiomed.2018.09.014. [DOI] [PubMed] [Google Scholar]
- 57.Chu Y., et al. Analyzing pre-symptomatic tissue to gain insights into the molecular and mechanistic origins of late-onset degenerative trinucleotide repeat disease. Nucleic Acids Res. Jul 9 2020;48(12):6740–6758. doi: 10.1093/nar/gkaa422. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 58.Dobin A., et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics. Jan 1 2013;29(1):15–21. doi: 10.1093/bioinformatics/bts635. (in eng) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 59.Robinson M.D., McCarthy D.J., Smyth G.K. edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics. Jan 1 2010;26(1):139–140. doi: 10.1093/bioinformatics/btp616. (in eng) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 60.Zhou N., Bao J. FerrDb: a manually curated resource for regulators and markers of ferroptosis and ferroptosis-disease associations. Database : the journal of biological databases and curation. Jan 1 2020;2020 doi: 10.1093/database/baaa021. (in eng) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 61.Rinkoski T.A., et al. Characterization of a dual media system for culturing primary normal and Fuchs endothelial corneal dystrophy (FECD) endothelial cells. PLoS One. 2021;16(9) doi: 10.1371/journal.pone.0258006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 62.Wisniewski J.R., Zougman A., Nagaraj N., Mann M. Universal sample preparation method for proteome analysis. Nat. Methods. May 2009;6(5):359–362. doi: 10.1038/nmeth.1322. [DOI] [PubMed] [Google Scholar]
- 63.Rappsilber J., Mann M., Ishihama Y. Protocol for micro-purification, enrichment, pre-fractionation and storage of peptides for proteomics using StageTips. Nat. Protoc. 2007;2(8):1896–1906. doi: 10.1038/nprot.2007.261. [DOI] [PubMed] [Google Scholar]
- 64.Chambers M.C., et al. A cross-platform toolkit for mass spectrometry and proteomics. Nat. Biotechnol. Oct 2012;30(10):918–920. doi: 10.1038/nbt.2377. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 65.Sturm M., et al. OpenMS - an open-source software framework for mass spectrometry. BMC Bioinf. Mar 26 2008;9:163. doi: 10.1186/1471-2105-9-163. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 66.Craig R., Beavis R.C. TANDEM: matching proteins with tandem mass spectra. Bioinformatics. Jun 12 2004;20(9):1466–1467. doi: 10.1093/bioinformatics/bth092. [DOI] [PubMed] [Google Scholar]
- 67.Geer L.Y., et al. Open mass spectrometry search algorithm. J. Proteome Res. Sep-Oct 2004;3(5):958–964. doi: 10.1021/pr0499491. [DOI] [PubMed] [Google Scholar]
- 68.Scott J. Dixon, et al. Ferroptosis: an iron-dependent form of nonapoptotic cell death. Cell. 2012/05/25/2012;149(5):1060–1072. doi: 10.1016/j.cell.2012.03.042. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 69.Kerins M.J., Ooi A. The roles of NRF2 in modulating cellular iron homeostasis. Antioxid Redox Signal. Dec 10 2018;29(17):1756–1773. doi: 10.1089/ars.2017.7176. (in eng) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 70.Fortenbach C.R., et al. Metabolic and proteomic indications of diabetes progression in human aqueous humor. PLoS One. 2023;18(1) doi: 10.1371/journal.pone.0280491. (in eng) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 71.Gao M., Monian P., Quadri N., Ramasamy R., Jiang X. Glutaminolysis and transferrin regulate ferroptosis. Mol. Cell. 2015;59(2):298–308. doi: 10.1016/j.molcel.2015.06.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 72.Jurkunas U.V. Fuchs endothelial corneal dystrophy through the prism of oxidative stress. Cornea. Nov 2018;37(Suppl 1):S50–s54. doi: 10.1097/ico.0000000000001775. (in eng) [DOI] [PubMed] [Google Scholar]
- 73.Liu J., Kuang F., Kroemer G., Klionsky D.J., Kang R., Tang D. Autophagy-dependent ferroptosis: machinery and regulation. Cell Chem. Biol. Apr 16 2020;27(4):420–435. doi: 10.1016/j.chembiol.2020.02.005. (in eng) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 74.Levi S., et al. A human mitochondrial ferritin encoded by an intronless gene. J. Biol. Chem. 2001/01/01/2001;276(27):24437–24440. doi: 10.1074/jbc.C100141200. [DOI] [PubMed] [Google Scholar]
- 75.Yang H., et al. Mitochondrial ferritin in neurodegenerative diseases. Neurosci. Res. Sep-Oct 2013;77(1–2):1–7. doi: 10.1016/j.neures.2013.07.005. (in eng) [DOI] [PubMed] [Google Scholar]
- 76.Gao G., Chang Y.-Z. Mitochondrial ferritin in the regulation of brain iron homeostasis and neurodegenerative diseases. Front. Pharmacol., Review. 2014-February-17 2014;5(19) doi: 10.3389/fphar.2014.00019. (in English) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 77.Sui X., et al. RSL3 drives ferroptosis through GPX4 inactivation and ROS production in colorectal cancer. Front. Pharmacol. 2018;9 doi: 10.3389/fphar.2018.01371. (in eng) 1371-1371. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 78.Bai Y., et al. Lipid storage and lipophagy regulates ferroptosis. Biochem. Biophys. Res. Commun. 2019/01/22/2019;508(4):997–1003. doi: 10.1016/j.bbrc.2018.12.039. [DOI] [PubMed] [Google Scholar]
- 79.Riegman M., et al. Ferroptosis occurs through an osmotic mechanism and propagates independently of cell rupture. Nat. Cell Biol. 2020;22(9):1042–1048. doi: 10.1038/s41556-020-0565-1. (in eng) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 80.Tyrrell R.M. Activation of mammalian gene expression by the UV component of sunlight--from models to reality. Bioessays : news and reviews in molecular, cellular and developmental biology. Feb 1996;18(2):139–148. doi: 10.1002/bies.950180210. (in eng) [DOI] [PubMed] [Google Scholar]
- 81.Pourzand C., Watkin R.D., Brown J.E., Tyrrell R.M. Ultraviolet A radiation induces immediate release of iron in human primary skin fibroblasts: the role of ferritin. Proceedings of the National Academy of Sciences of the United States of America. 1999;96(12):6751–6756. doi: 10.1073/pnas.96.12.6751. (in eng) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 82.Valerio H.P., Ravagnani F.G., Yaya Candela A.P., Dias Carvalho da Costa B., Ronsein G.E., Di Mascio P. Spatial proteomics reveals subcellular reorganization in human keratinocytes exposed to UVA light. iScience. 2022/04/15/2022;25(4) doi: 10.1016/j.isci.2022.104093. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 83.Jugé R., Breugnot J., Da Silva C., Bordes S., Closs B., Aouacheria A. Quantification and characterization of UVB-induced mitochondrial fragmentation in normal primary human keratinocytes. Sci. Rep. 2016/10/12 2016;6(1) doi: 10.1038/srep35065. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 84.Bersuker K., et al. The CoQ oxidoreductase FSP1 acts parallel to GPX4 to inhibit ferroptosis. Nature. Nov 2019;575(7784):688–692. doi: 10.1038/s41586-019-1705-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 85.Fan Z., et al. Nrf2-Keap1 pathway promotes cell proliferation and diminishes ferroptosis. Oncogenesis. Aug 14 2017;6(8):e371. doi: 10.1038/oncsis.2017.65. (in eng) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 86.Hou W., et al. Autophagy promotes ferroptosis by degradation of ferritin. Autophagy. 2016/08/02 2016;12(8):1425–1428. doi: 10.1080/15548627.2016.1187366. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 87.Xu C., et al. The glutathione peroxidase Gpx4 prevents lipid peroxidation and ferroptosis to sustain Treg cell activation and suppression of antitumor immunity. Cell Rep. Jun 15 2021;35(11) doi: 10.1016/j.celrep.2021.109235. (in eng) [DOI] [PubMed] [Google Scholar]
- 88.Geng N., et al. Knockdown of ferroportin accelerates erastin-induced ferroptosis in neuroblastoma cells. Eur. Rev. Med. Pharmacol. Sci. Jun 2018;22(12):3826–3836. doi: 10.26355/eurrev_201806_15267. (in eng) [DOI] [PubMed] [Google Scholar]
- 89.Ward D.M., Kaplan J. Ferroportin-mediated iron transport: expression and regulation. Biochim. Biophys. Acta. 2012;1823(9):1426–1433. doi: 10.1016/j.bbamcr.2012.03.004. (in eng) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 90.Harada N., et al. Nrf2 regulates ferroportin 1-mediated iron efflux and counteracts lipopolysaccharide-induced ferroportin 1 mRNA suppression in macrophages. Arch. Biochem. Biophys. Apr 1 2011;508(1):101–109. doi: 10.1016/j.abb.2011.02.001. (in eng) [DOI] [PubMed] [Google Scholar]
- 91.Fisenko N.V., Trufanov S.V., Avetisov K.S., Vtorushina V.V., Subbot A.M. Evaluation of aqueous cytokine levels in eyes with Fuchs endothelial corneal dystrophy and bullous keratopathy. Vestn. Oftalmol. 2021;137(3):13–18. doi: 10.17116/oftalma202113703113. Opredelenie urovnya tsitokinov vo vnutriglaznoi zhidkosti pri endotelial'noi distrofii rogovitsy Fuksa i bulleznoi keratopatii. [DOI] [PubMed] [Google Scholar]
- 92.Kuot A., et al. Differential gene expression analysis of corneal endothelium indicates involvement of phagocytic activity in Fuchs' endothelial corneal dystrophy. Exp. Eye Res. Sep 2021;210 doi: 10.1016/j.exer.2021.108692. [DOI] [PubMed] [Google Scholar]
- 93.Mao C., et al. DHODH-mediated ferroptosis defence is a targetable vulnerability in cancer. Nature. May 2021;593(7860):586–590. doi: 10.1038/s41586-021-03539-7. (in eng) [DOI] [PMC free article] [PubMed] [Google Scholar]
- 94.Gain P., et al. Global survey of corneal transplantation and eye banking. JAMA Ophthalmol. Feb 2016;134(2):167–173. doi: 10.1001/jamaophthalmol.2015.4776. (in eng) [DOI] [PubMed] [Google Scholar]
- 95.Vile G.F., Tyrrell R.M. UVA radiation-induced oxidative damage to lipids and proteins in vitro and in human skin fibroblasts is dependent on iron and singlet oxygen. Free Radic. Biol. Med. Apr 1995;18(4):721–730. doi: 10.1016/0891-5849(94)00192-m. [DOI] [PubMed] [Google Scholar]
- 96.Morliere P., Moysan A., Santus R., Huppe G., Maziere J.C., Dubertret L. UVA-induced lipid peroxidation in cultured human fibroblasts. Biochim. Biophys. Acta. Jul 30 1991;1084(3):261–268. doi: 10.1016/0005-2760(91)90068-s. [DOI] [PubMed] [Google Scholar]
- 97.Punnonen K., Jansen C.T., Puntala A., Ahotupa M. Effects of in vitro UVA irradiation and PUVA treatment on membrane fatty acids and activities of antioxidant enzymes in human keratinocytes. J. Invest. Dermatol. Feb 1991;96(2):255–259. doi: 10.1111/1523-1747.ep12462271. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.