Activation of thioredoxin system, HIF-1a and pro-ferroptotic pathways by oxidative stress in human iPSC-derived retinal pigment epithelium cells

Cell culture

The hiPSC-derived RPE cells (donor sex: male, population: Caucasian) were received at passage 2 from Phenocell (Grasse, France) and were cultured in RPE medium containing 70% Dulbecco’s modified Eagle’s medium, high glucose, 30% Ham’s F12 Nutrient Mix, 2% B-27 supplement, and 1% antibiotic-antimycotic, all from Gibco, Thermo Fisher Scientific (Waltham, MA). Cells were used at passages 3 to 4 for experiments. Passaging of hiPSC-RPE cells was performed using Accumax solution (Sigma-Aldrich, St. Louis, MO), and cells were seeded on Matrigel (Corning, Corning, NY)–coated permeable Transwell cell culture inserts (Corning) and 96-well plates (µClear; Greiner Bio-One, Kremsmünster, Austria) at a density of 100,000 cells/cm². The cells were cultured on inserts for 28 days and in 96-well plates for 22 days at 37 °C, 5% CO₂. Cells tested negative for mycoplasma using MycoAlert (Lonza, Basel, Switzerland).

Immunocytochemistry

Cells cultured on inserts were washed with Dulbecco’s phosphate-buffered saline (DPBS; Gibco, Thermo Fisher Scientific), fixed with methanol for 10 min, washed again, and blocked in blocking buffer (5% goat serum in DPBS) for 20 min. The membranes containing the cells were cut out from the inserts and transferred to a well plate. Cells were incubated in different primary antibodies, including anti–zonula occludens 1 (1:100; Thermo Fisher Scientific), anti-occludin (1:100, Thermo Fisher Scientific), anti-RPE65 (1:100; Novus Biologicals, Bio-Techne, Minneapolis, MN), and anti-ARL13B (1:100; Proteintech, Manchester, UK) for 90 min; washed with washing solution (0.5% goat serum in DPBS); and incubated in secondary antibody for 1 h. Secondary antibodies were goat anti-rabbit or anti-mouse IgG Alexa Fluor 488 (1:1,000; Thermo Fisher Scientific), and for ARL13B staining, the secondary antibody was excess Fab fragment goat anti-rabbit Alexa Fluor 594 (1:100; Jackson ImmunoResearch Europe Ltd, Cambridgeshire, UK). After washing, the cell nuclei were stained with 0.1 µg/ml 4′,6-diamidino-2-phenylindole (DAPI; Sigma-Aldrich) for 10 min. The cells were mounted on microscope slides using FluoroShield (Sigma-Aldrich). Images were acquired using a fluorescence microscope (Leica Microsystems GmbH, Wetzlar, Germany).

Transmission electron microscopy

Cells cultured on inserts were washed with DPBS, fixed with 2% paraformaldehyde and 0.5% glutaraldehyde for 10 min, and washed with 0.1 M phosphate buffer (pH 7.4). Cells were postfixed in 1% osmium tetroxide (Electron Microscopy Sciences, Hatfield, PA) for 1 h, washed, and dehydrated in a series of ethanol (50%, 70%, 96%, and absolute ethanol). An epoxy resin embedding kit (Ladd Research, Williston, VT) was used for cell infiltration and embedding according to the manufacturer’s instructions. Ultrathin sections (70 nm) were cut and stained with uranyl acetate and lead citrate (both from Electron Microscopy Sciences). Images were obtained with a JEM-2100F transmission electron microscopy (TEM) microscope (Jeol, Tokyo, Japan).

Transepithelial electrical resistance assessment

Culture medium was removed from the inserts, and cells were washed with warm Hank’s Balanced Salt Solution (HBSS; Gibco, Thermo Fisher Scientific) with 25 mM HEPES (Sigma-Aldrich). After washing, the cells were balanced with the same buffer for 30 min at 37 °C, and TEER values were measured using an EVOM2 Epithelial Volt/Ohm Meter (World Precision Instruments, Sarasota, FL). After the permeability experiments, test solutions were removed, fresh buffer was added in the chambers, and TEER values were measured as described above.

Permeability testing

The permeability tests were performed in inserts from the basolateral to apical direction (i.e., representing transport from choroid to retina). Reference compounds were diluted in HBSS-25 mM HEPES solution, 6-carboxyfluorescein (100 µM), and rhodamine B (100 µM) from Sigma-Aldrich and placed in the donor chambers. Fresh buffer was added to the receiver chambers. The samples were taken from the receiver chambers at 10, 20, 30, 40, 50, and 60 min and from the donor chambers at 0, 30, and 60 min. The loss of volume in the receiver chambers was replaced with equal volumes of fresh buffer. The fluorescence intensity of the samples was measured using a Cytation 3 multimode reader (BioTek Instruments, Winooski, VT), and sample concentrations were determined using standard curves. Three independent experiments in duplicate were performed (n = 6).

Induction of oxidative stress

Oxidative stress was induced in hiPSC-RPE cells grown in 96-well plates with NaIO₃ in concentrations of 0 to 20 mM and tert-butyl hydroperoxide (tBHP) in concentrations of 0 to 1.2 mM (both from Sigma-Aldrich) at 37 °C for 24 h. NaIO₃ was diluted in RPE culture medium with B-27 supplement without antioxidants (Thermo Fisher Scientific).

Antioxidant treatment

To evaluate antioxidant-mediated suppression of NaIO₃-induced effects, cells were treated with manganese (III) tetrakis(1-methyl-4-pyridyl)porphyrin (Mn-TM-2-PyP, 0.005%), a cell-permeable antioxidant with broad redox activity, for 24 h before and during NaIO₃ exposure. The experiment involving antioxidant was conducted only once for demonstrative purposes, and the results should be interpreted as preliminary and considered directional rather than conclusive. Further replication is required to confirm these findings.

Lactate dehydrogenase and VEGFA release assays

Culture medium was sampled to assess lactate dehydrogenase (LDH) and VEGFA release from the hiPSC-RPE cells. LDH was measured as previously described [27]. VEGFA concentrations from the cell supernatants were measured using the Human VEGFA Quantikine ELISA Kit according to the manufacturer’s instructions (R&D Systems, Bio-Techne, Minneapolis, MN). Three independent experiments (n = 3) were performed with four to eight technical replicates.

Resazurin cell viability and ROS assays

The reduction of resazurin to fluorescent resorufin was used to assess cell viability. Resazurin (Sigma-Aldrich) 0.01 mg/ml in HBSS was added to the cells, which were then incubated at 37 °C for 1 h. The formation of ROS was measured with a general oxidative stress indicator, CM-H₂DCFDA (Thermo Fisher Scientific), which was added to the cells at 5 µM in HBSS and incubated at 37 °C for 1 h (loading), after which the cells were incubated in HBSS at 37 °C for an additional 30 min. Fluorescence intensities were measured using the Cytation 3 reader. Three independent experiments (n = 3) were performed with two to four technical replicates. After the fluorescent measurements, cells were washed with DPBS and lysed in TRIzol reagent (Thermo Fisher Scientific).

RNA purification

NaIO₃-induced hiPSC-RPE cells were lysed in TRIzol reagent, and cells from 2 wells were pooled. Phase separation and RNA isolation were performed from the colorless upper aqueous phase with a PureLink RNA Mini Kit (Thermo Fisher Scientific) according to the manufacturer’s instructions. Optional on-column PureLink DNase treatment was performed to minimize DNA contents in purified RNA samples. RNA concentration and purity were determined with a NanoDrop One^C Microvolume UV-Vis Spectrophotometer (Thermo Fisher Scientific).

Complementary DNA synthesis and real-time quantitative polymerase chain reaction amplification

Complementary DNA (cDNA) synthesis was performed with the High-Capacity RNA-to-cDNA Kit (Thermo Fisher Scientific) according to manufacturer’s instructions using 0.2 µg total RNA from cells. Then, 5 ng cDNA samples were used for target (carboxyfluorescein [FAM]-labeled) and endogenous control (2′-chloro-7′phenyl-1,4-dichloro-6-carboxy-fluorescein [VIC]-labeled) gene quantitative polymerase chain reaction (qPCR) amplification with five different TaqMan assays: vascular endothelial growth factor A (VEGFA, Hs00900055_m1), hypoxia-inducible factor 1 alpha (HIF1A, Hs00153153_m1), thioredoxin (TXN, Hs00828652_m1), thioredoxin-interacting protein (TXNIP, Hs00197750_m1), ChaC glutathione-specific gamma-glutamylcyclotransferase 1 (CHAC1, Hs00225520_m1), cystine/glutamate antiporter, solute carrier family 7 member 11 (SLC7A11, Hs00921938_m1), and glyceraldehyde-3-phosphate dehydrogenase (GAPDH, Hs99999905_m1, endogenous control; all from Thermo Fisher Scientific), using the TaqMan Fast Advanced Master Mix in QuantStudio 3 Real-Time PCR System (Thermo Fisher Scientific) according to the manufacturer’s instructions with an elongated initial DNA denaturation time (95 °C for 2 min). Two separate experiments were conducted, including six technical replicates. Each technical replicate was analyzed in triplicate. qPCR data were normalized to GAPDH and analyzed with the 2^–ΔΔCT method [28] for relative gene expression analysis.

Statistical analyses

For the permeability testing, the apparent permeability coefficient (P_app, cm/s) was calculated according to the following equation: P_app = (∆Qr / ∆t) / (C_d × A), where ∆Qr/∆t is the slope of the linear region of the cumulative amount of the compound in the receiver chamber versus time plot, C_d is the average concentration of the compound in the donor chamber during the experiment, and A is the surface area of the insert (1.12 cm²). Data are presented as mean ± standard deviation (SD).

The relative fluorescence unit (RFU) of background was measured and deducted from all treated and control conditions, and cell viability (% of control) was calculated with the following equation: RFU_treated/RFU_control × 100. The four-parameter dose–response curve was fitted using GraphPad Prism (version 10.4.0; Dotmatics, Boston, MA) to obtain the half-maximal inhibitory concentration (IC₅₀), defined as the concentration of NaIO₃ that reduces cell viability to 50%. LDH and ROS data were normalized to the control condition (0 mM NaIO₃). LDH data were analyzed as a four-parameter dose–response curve, and the half-maximal effective concentration (EC₅₀) for the LDH assay was calculated in GraphPad Prism software. Data are presented as mean ± SD.

Relative qPCR data were normalized to the control condition (0 mM NaIO₃). Quantitative data were analyzed using descriptive statistics and presented as mean ± SD. Statistical analyses were performed using GraphPad Prism, and differences were considered statistically significant at the multiplicity-adjusted p < 0.05 level. For normally distributed data, comparisons between the groups were determined using one-way analysis of variance, followed by Dunnett’s multiple comparison test or two-way analysis of variance with Šidák’s multiple comparison test.

Characteristics of native human RPE phenotype were expressed in hiPSC-RPE cells

The fluorescence images of zonula occludens 1 and occludin revealed that both proteins were strongly expressed at cell–cell junctions, and the typical polygonal shape of the cells was also clearly visible (Figure 1A, B). TEM images showed that the cells possessed several typical characteristics of native RPE cells, such as apical microvilli and melanosomes at all maturation stages, I to IV (Figure 1C, D). Interestingly, we also observed that occludin consistently stained a dot-like region in the middle of each hiPSC-RPE cell (Figure 1B and Figure 2A). Previous studies have observed occludin localization in centrosomes of primary cilia during cortex development [29]. Here, we explored whether occludin (Figure 2A) is costained with primary cilia marker ARL13B (Figure 2B). Indeed, double immunostaining of hiPSC-RPE cells and DAPI counterstaining revealed colocalization of occludin and ARL13B (Figure 2C). Lastly, we immunostained the cells against RPE-specific 65 kDa protein (RPE65), which was also expressed in the hiPSC-RPE cells (Figure 2D). Taken together, these maturity markers confirm the full maturation of the hiPSC-RPE cells grown in inserts [30].

Melanosome density and maturation in hiPSC-RPE cells

The density and maturation level of melanosomes were evaluated from TEM images (Figure 3A). In total, melanosomes were counted from 24 hiPSC-RPE cells. Similarly, as reported by others [31,32], we found that mature melanosomes of stages III and IV were more predominant in hiPSC-RPE cells than immature melanosomes of stages I and II (Figure 3A,B).

hiPSC-RPE cells presented tight junction functionality and monolayer tightness

The TEER values of hiPSC-RPE cells were measured to confirm the integrity of the cell monolayer grown for 28 days in inserts. The results showed high values both before (481 ± 36 Ω × cm²) and after the permeability testing (549 ± 34 Ω × cm²). The permeability testing of the reference compounds from the basolateral to apical direction showed that the reference compound with high permeability, rhodamine B, had the high P_app value (18.9 ± 7.7 × 10⁻⁶ cm/s, n = 6), while the low-permeability marker, 6-carboxyfluorescein (2.1 ± 1.2 × 10⁻⁶ cm/s, n = 6), had a significantly lower P_app value, indicating a strong epithelial barrier with good discrimination between low- and high-permeability values.

NaIO₃-induced oxidative stress decreased cell viability, induced ROS and VEGFA production, and promoted antioxidative, proangiogenic, and pro-ferroptotic gene expression in hiPSC-RPE cells

Cell viability decreased when the hiPSC-RPE cells were treated with NaIO₃ for 24 h. A decrease in resazurin was observed at 5 mM NaIO₃ and reached its maximum at 20 mM, generating an IC₅₀ of 9.4 mM (Figure 4A). Meanwhile, LDH levels started to rise with 3 mM NaIO₃ and rose steadily as the concentration increased to 10 mM NaIO₃, after which the LDH levels stayed stable despite increasing concentrations, resulting in an IC₅₀ of 4.1 mM (Figure 4B). This indicates that NaIO₃ caused cell membrane damage first before affecting intracellular metabolism.

Interestingly, ROS formation was remarkably increased, even with the lowest NaIO₃ concentrations (0.1–2 mM; Figure 4A), while viability indicators did not show decreased cell viability. The ROS levels gradually decreased from 2 mM NaIO₃ onward, coinciding with the cell membrane damage indicated by LDH release (Figure 4B). In addition, an increase in VEGFA levels could already be seen with 1 mM NaIO₃ when the ROS levels peaked, and the VEGFA levels rose further to approximately 4 mM NaIO₃ (Figure 4B).

We sought to validate the principal cellular responses—cell viability, ROS generation, and the release of VEGFA—in hiPSC-RPE cells using an alternative oxidative agent, tBHP, which, although commonly used to induce oxidative stress in cellular models, is not specific to RPE cells. Compared with NaIO₃, tBHP produced a markedly steeper dose–response curve. The IC₅₀ for resazurin reduction was 0.59 mM, while the EC₅₀ for LDH release was 0.55 mM, indicating that tBHP induces intracellular metabolic impairment and cell membrane damage at approximately the same tBHP concentration (Figure 4C, D).

Interestingly, tBHP did not induce a comparable peak in ROS formation at lower concentrations of oxidative agent, in contrast to NaIO₃. Moreover, overall ROS levels were lower following tBHP induction (Figure 4C) than those observed with NaIO₃ (Figure 4A). Furthermore, VEGFA levels were elevated (Figure 4D), but the increase was less pronounced compared with NaIO₃ induction (Figure 4B).

Gene expression profiles of VEGFA, HIF1A, TXN, TXNIP, CHAC1, and SLC7A11 were generated from NaIO₃-treated cells (Figure 5). The results revealed that VEGFA, HIF1A, TXN, CHAC1, and SLC7A11 gene expression were significantly upregulated (p < 0.05), and TXNIP gene expression was significantly downregulated (p < 0.05) in cells treated with 2 to 6 mM NaIO₃ compared with untreated cells. Upregulation of the TXN gene, encoding thioredoxin, may lead to a reduction in ROS formation in hiPSC-RPE cells at 2 mM and higher concentrations of NaIO₃ (Figure 4A). Downregulation of the TXNIP gene (Figure 5D) may lead to an increase in VEGFA (Figure 4B). Oxidative stress induced upregulation of the ferroptosis mediator CHAC1, leading to enhanced glutathione degradation, mitochondrial dysfunction, and disruption of redox homeostasis (Figure 5E). Concurrently, increased expression of the cystine/glutamate antiporter (SLC7A11) gene at 2 mM NaIO₃ (Figure 5F) promoted cystine uptake and glutathione synthesis, which may represent an early adaptive response to mitigate oxidative stress and limit ferroptotic cell death.

Mn-TM-2-PyP–mediated antioxidant treatment suppressed NaIO₃-induced cytotoxicity in cell viability assays. The effect was less pronounced in the resazurin reduction-based cell viability assay, with a modest effect in the IC₅₀ value shift from 9.0 to 11.5 mM (Figure 6A). In contrast, LDH release was markedly reduced, with a pronounced rightward shift in the EC₅₀ from 5.1 to 11.0 mM (Figure 6B). Since gene expression changes were most pronounced at NaIO₃ concentrations of 4 to 6 mM across all genes tested (Figure 5), antioxidant effects were evaluated at these levels (Figure 6C–H). Antioxidant treatment consistently attenuated gene upregulation, reaching statistical significance for VEGFA and TXN at 4 and 5 mM NaIO₃ (Figure 6C, E). Notably, TXNIP gene expression was higher in antioxidant-treated cells exposed to 5 mM NaIO₃ than in untreated controls (Figure 6F). These results suggest that antioxidant treatment helps to restore gene expression toward baseline in cells experiencing extensive oxidative stress. However, antioxidant-containing data should be interpreted as preliminary, and further replication is required to confirm these findings.