Nitric Oxide May Adversely Affect the Metabolism and Viability of Retinal Organoids Derived from Patients with Leber Hereditary Optic Neuropathy

Abstract

Leber hereditary optic neuropathy (LHON) is a bilateral optic neuropathy associated with mitochondrial DNA (mtDNA) mutations characterized by parapapillary telangiectasia during the acute phase. However, its precise mechanism remains unclear. This study evaluated the effects of nitric oxide (NO) on retinal organoids (ROs) generated from induced pluripotent stem cells derived from patients with LHON. Established induced pluripotent stem cells from three patients with the m.11778G>A mutation (patient group) and three healthy individuals (control group) were differentiated into ROs. Changes in cell death ratios, mtDNA copy number, and metabolite profiles in the ROs following exposure to sodium nitroprusside (SNP), which was an NO donor, were compared between the two groups. At baseline, terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling–positive cell ratios did not differ significantly, whereas the mtDNA copy number was significantly higher in the patient group. SNP exposure significantly increased the proportion of terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling–positive cells in the patient group but did not affect the mtDNA copy number. Relative concentrations of metabolites, including taurine and γ-aminobutyric acid, were initially reduced in the patient group, but increased following SNP exposure. These findings suggest that NO may promote retinal cell death and disrupt metabolite profiles in ROs derived from patients with LHON.

Leber hereditary optic neuropathy (LHON) is a painless, acute, or subacute bilateral optic neuropathy definitively associated with mitochondrial DNA (mtDNA) mutations and potentially with chromosomal DNA mutations.¹ It primarily affects young males, with a prevalence of approximately 1:50,000.^2,3 The main symptoms of LHON are decreased visual acuity and central scotoma. They severely interfere with daily life activities and markedly impair quality of life.³

Pathogenic mtDNA mutations in one of the subunits of oxidative phosphorylation electron transfer complex I are believed to reduce ATP production and/or increase reactive oxygen species levels, leading to retinal ganglion cell (RGC) apoptosis. However, additional genetic, epigenetic, or environmental triggers are likely required. These include mitophagy,⁴ reproductive hormones,⁵ smoking,⁶ and various other oxidative stresses.⁷ However, the precise mechanism of RGC death in LHON remains unclear. Moreover, LHON also increases both mortality and comorbidity.⁸ In addition, pathogenetic changes may also occur in photoreceptors.⁹ These findings imply that studying RGC pathology alone will not lead to a complete understanding or effective treatment for LHON, and at least, the pathology needs to be evaluated in the whole retina.

One of the clinical hallmarks of the acute phase of LHON is parapapillary microvascular telangiectasia, which precedes decreased visual function. It is a diagnostic sign that makes clinicians differentiate LHON from other causes of optic nerve diseases, such as optic neuritis and anterior ischemic optic neuropathy, because of the lack of leakage from the dilated capillary on fluorescence angiography. Although the mechanisms underlying this unique telangiectatic change remain unknown, it is reasonable to speculate that nitric oxide (NO) is involved in this event because NO not only dilates capillaries without tight junction barrier function but also exerts a neurotrophic and neurotoxic effect in a dose-dependent manner.¹⁰ NO also regulates respiration via the mitochondrial electron transfer complex and contributes to mitochondrial biogenesis through cGMP or peroxisome proliferator–activated receptor γ coactivator 1-α.¹¹ These observations led to the proposal of the following hypothesis: excessive NO stress in the juxtapapillary region of patients with LHON contributes to RGC death. As an initial step to test this hypothesis, the effects of NO on cell survival, mtDNA copy number, and metabolite profiles were assessed in retinal organoids (ROs) derived from the induced pluripotent stem (iPS) cells of patients with LHON, comparing them with those from healthy controls.

Materials and Methods

Establishment of iPS Cells

This study followed the tenets of the Declaration of Helsinki and was approved by the institutional review board of the Kobe University Graduate School of Medicine (Kobe, Japan; authorization number: B230085). Three control iPS cell lines (control group: controls 1 to 3) and three patient-derived iPS cell lines (patient group: patients 1 to 3) were prepared. Control 1 and all patient lines were established at the Laboratory for Retinal Regeneration, RIKEN Center for Biosystems Dynamics Research, and deposited to the Kobe University Graduate School of Medicine via a material transfer agreement (agreement number: MTAH29-119). All patient lines originated from male patients aged in their 20s to 40s, who harbored m.11778G>A mutation—the major mutation in LHON. Informed consent was obtained, and 40 mL of peripheral blood was collected. Each iPS cell line was generated by electroporating episomal vectors into mononuclear cells, as previously described.^12,13 Control 2 and 3 lines were deposited by the RIKEN BioResource Research Center, Kyoto University (Kyoto, Japan). All controls (including control 1) were derived from male donors aged in their 20s to 40s. Detailed information on the origin and quality of controls 2 and 3 is available online (https://cellbank.brc.riken.jp/cell_bank/CellInfo/?cellNo=HPS2496&lang=En and https://cellbank.brc.riken.jp/cell_bank/CellInfo/?cellNo=HPS3344&lang=En, last accessed April 25, 2025). Both lines were also established from peripheral blood via electroporation of exogenous genes (Oct3/4, Sox2, Klf4, RLF, Lin28A, mp53DD, and EBNA1BP2).

Differentiation of Three-Dimensional ROs

iPS cells were maintained and differentiated into ROs using the serum-free floating culture of embryoid body–like aggregates with quick aggregation method.¹⁴ Briefly, iPS cell colonies were maintained in StemFit AK02N with supplement C (AJINOMOTO, Tokyo, Japan). One day before the start of differentiation, colonies were treated with 5 μmol/L SB431542 (Merck KGaA, Darmstadt, Germany) and 300 nmol/L smoothened agonist (Enzo, Farmingdale, NY). iPS cells were then dissociated into 96-well V-bottom plates (Sumitomo Bakelite, Tokyo, Japan) at 12,000 cells/well. Cultures were maintained for 6 days in differentiation medium (1:1 mixture of Iscove's modified Dulbecco's medium and F12 medium; Thermo Fisher Scientific, Waltham, MA) supplemented with 10% knockout serum replacement, 1-thioglycerol, lipid emulsion, 0.5% bovine serum albumin, 100 U/mL penicillin, and 100 μg/mL streptomycin. On day 6, 1.5 nmol/L human bone morphogenetic protein 4 (R&D Systems, Minneapolis, MN) was added to each well. Half of the medium was replaced every 3 days. On day 18 or 19, ROs were transferred to 90-mm low-cell binding dishes (Sumitomo Bakelite) containing Dulbecco's modified Eagle's medium: Nutrient Mixture F-12 (Dulbecco’s modified Eagle’s medium/F12), GlutaMAX Supplement (Thermo Fisher Scientific), 1% N2 supplement (Thermo Fisher Scientific), vascular endothelial growth factor receptor/fibroblast growth factor receptor inhibitor (SU5402; Merck KGaA), and 3 mmol/L glycogen synthase kinase-3 inhibitor (CHIR99021; Cayman Chemical, Ann Arbor, MI). On day 21, the medium was changed to Dulbecco’s modified Eagle’s medium/F12 GlutaMAX supplemented with 10% fetal bovine serum, 1% N2 supplement, 100 nmol/L retinoic acid (Merck KGaA), 50 μmol/L taurine (Merck KGaA), 100 U/mL penicillin, 100 μg/mL streptomycin, and 0.25 μg/mL amphotericin B. The medium was replaced twice per week before each experiment.

Immunohistochemistry for Confirming Differentiation

ROs were fixed with 3% paraformaldehyde/phosphate-buffered saline (PBS) containing 7.5% sucrose at 4°C, washed twice in 7.5% sucrose/PBS for 5 minutes, cryoprotected in 15% followed by 30% sucrose, embedded in OCT compound (Sakura Finetek Japan, Tokyo, Japan), and cryopreserved.

Frozen sections were blocked in 5% horse serum, incubated overnight with primary antibodies (Table 1) in 1% horse serum, followed by secondary antibodies conjugated with Alexa Fluor 488 or 546, after which they were counterstained with DAPI and then imaged using a confocal microscope (LSM700; Carl Zeiss, Oberkochen, Germany).

Table 1.

List of Primary Antibodies Used in the Study

Antibody no.	Antibody	Supplier	Catalog no.	Dilution
1	AFP	R&D Systems	MAB1369	1:100
2	α-SMA	Invitrogen (Waltham, MA)	14-9760-82	1:1000
3	Tuj-1	BioLegend (San Diego, CA)	801213	1:1000
4	Oct3/4	Santa Cruz Biotechnology (Dallas, TX)	sc5279	1:200
5	NANOG	Millipore (Burlington, MA)	MABD24	1:500
6	TRA-1-60	Millipore	MAB4360	1:200
7	SSEA-4	Millipore	MAB4304	1:200
8	Rx	Takara Bio (Shiga, Japan)	M229	1:500
9	CRX	Takara Bio	M231	1:500
10	MATH5/ATOH7	Abcam (Cambridge, UK)	ab229245	1:500
11	NeuN	Abcam	ab104224	1:500

AFP, alpha-fetoprotein; α-SMA, α-smooth muscle actin; CRX, cone-rod homeobox protein; Rx, retinal homeobox protein; Tuj-1, neuron-specific class III beta-tubulin.

iPS cell colonies and embryoid bodies were stained with undifferentiated and pluripotent markers (Supplemental Figure S1), respectively. Colonies and embryoid bodies were fixed with 1.5% or 3.0% paraformaldehyde/PBS at 4°C and washed with PBS. For OCT3/4 and NANOG staining, samples were incubated with 0.25% PBS with Tween 20. Primary and secondary antibody procedures followed those used for organoid immunostaining.

Exposure of ROs to SNP

On approximately day 35, ROs were exposed to sodium nitroprusside dihydrate (Merck KGaA) for 24 or 72 hours. The exposure was conducted under three conditions: i) a single 24-hour exposure to increasing concentrations (0, 0.1, 1.0, 5.0, and 10.0 mmol/L), ii) three cycles of 24-hour exposures to 1 mmol/L sodium nitroprusside (SNP) with 24-hour SNP-free intervals, and iii) a continuous 72-hour exposure to 1 mmol/L SNP.

Measurement of Relative mtDNA Copy Number

Three ROs were used for each SNP exposure condition. Genomic DNA was extracted using the DNeasy Blood and Tissue Kit (Qiagen, Hilden, Germany), following the manufacturer's instructions. The mtDNA copy number relative to nuclear DNA was measured using real-time PCR with a human mtDNA detection primer set [Human Mitochondrial DNA (mtDNA) Monitoring Primer Set; Takara Bio Inc., Shiga, Japan]. Briefly, the extracted genomic DNA was amplified with four primer pairs: mtDNA primer pairs ND1 and ND5, and nuclear DNA primer pairs SLCO2B1 and SERPINA1, to detect mtDNA and nuclear DNA. PCR cycling conditions included 95°C for 30 seconds, followed by 40 cycles of 95°C for 5 seconds and 60°C for 30 seconds. The difference in Ct values between ND1 and SLCO2B1 was defined as ΔCt1, and that between ND5 and SERPINA1 as ΔCt2. Relative mtDNA copy numbers were calculated as the averages of 2^ΔCt1 and 2^ΔCt2. Each experiment was independently repeated three times, with three technical replicates per sample.

Terminal Deoxynucleotidyl Transferase-Mediated dUTP Nick-End Labeling Staining

Terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) staining was performed using the In Situ Cell Death Detection Kit, Fluorescein (Roche, Basel, Switzerland), following the manufacturer's protocol. Briefly, sections were permeabilized with a mixture of 50 mL of MilliQ water, 500 μL of 10% sodium citrate dihydrate solution (50 mL of MilliQ water and 5 g of 10% sodium citrate), and 500 μL of 10% Triton-X for 2 minutes on ice. TUNEL labeling solution was prepared in a 1:9 ratio and applied to sections at 37°C for 1 hour.

TUNEL-positive areas were quantified using ImageJ software version 1.54 (NIH, Bethesda, MD; https://imagej.net/ij). Given the SNP-induced morphologic disruption in some organoids, the TUNEL-positive area was normalized to the retinal layer area (Supplemental Figure S2). Retinal regions were selected from organoid sections, and thresholds were set based on TUNEL staining. The TUNEL positivity rate was calculated from three organoids per condition, and mean values were used for analysis.

Gas Chromatography–Mass Spectrometry Analysis of ROs

Metabolomic profiling was conducted as previously described.^15,16 Low-molecular-weight metabolites were extracted by mixing frozen ROs with 250 μL of methanol-water-chloroform solvent mixture (2.5:1:1, v/v/v) containing 10 μL of aqueous sinapinic acid (0.5 mg/mL in distilled water; Sigma Aldrich, St. Louis, MO) as an internal standard. Samples were shaken at 1200 rpm and 37°C for 30 minutes and centrifuged for 3 minutes at 4°C at 22,000 × g. Subsequently, 200 μL of distilled water was added to 225 μL of supernatant, and the solution was centrifuged for 3 minutes at 4°C and 22,000 × g. Following this step, 250 μL of supernatant was lyophilized in a fresh tube using a freeze dryer.

The lyophilized samples were mixed with 20 μL of methoxyamine hydrochloride (20 mg/mL in pyridine; Sigma Aldrich), shaken at 1200 rpm for 90 minutes at 30°C to perform oximation, centrifuged at 22,000 × g for 5 minutes at 4°C after adding 10 μL of N-methyl-N-trimethylsilyl-trifluoroacetamide (GL Science, Tokyo, Japan) for derivatization, and incubated at 1200 rpm for 30 minutes at 37°C.

Gas chromatography–mass spectrometry analysis was conducted using a GCMS-QP2010 Ultra (Shimadzu Co, Kyoto, Japan) equipped with a fused silica capillary column (CP-SIL 8 CB low bleed/MS; 30 m × 0.25-mm inner diameter, 0.25-μm film thickness; Agilent Co, Palo Alto, CA) at the Integrated Center for Mass Spectrometry, Kobe University Graduate School of Medicine. The helium flow rate was 39.0 cm/second, and the front inlet temperature was set at 230°C. Column temperature was maintained at 80°C for 2 minutes, increased to 330°C at 15°C/minute, and held for 6 minutes. Transfer line and ion-source temperatures were set at 250°C and 200°C, respectively. Data acquisition included 20 scans across the 85 to 500 m/z mass range using advanced scanning speed protocol (Shimadzu Co).

Data processing, including peak detection, alignment, and metabolite identification, was performed using MS-DIAL software version 3.9.6 after converting raw files to netCDF format.¹⁷ Metabolite annotations were referenced against a library of 438 compounds. Ion peak heights were normalized to the peak height of sinapinic acid as the internal standard and adjusted for tissue weight.

Metabolite Analysis

Metabolite data exported from MS-DIAL in comma separated values (CSV) format were analyzed using MetaboAnalyst version 6.0 (https://www.metaboanalyst.ca/home.xhtml). Data normalization included log transformation and autoscaling. As all gas chromatography–mass spectrometry measurements were conducted on the same day, the inspection date block was not adjusted. Principal component analysis was used to achieve natural interactions between samples (grouping, clustering, and outlier detection) and quality control.¹⁶ To identify the 25 metabolites with the greatest intergroup variation, individual- and group-based hierarchical clustering analyses were performed.

Discriminative metabolites between groups were determined using partial least squares discriminant analysis with variable importance in projection scoring.^16,18 Metabolite alterations with variable importance in projection scores of >1 were considered differentially altered.¹⁹ Metabolite set enrichment analysis was performed using MetaboAnalyst software, where the Small Molecule Pathway Database (http://www.smpdb.ca, last accessed December 7, 2024) was chosen as a metabolite set library. Enrichment analysis was based on a global test approach²⁰ using a generalized linear model to compute a Q-stat for each metabolite set. The Q-stat is calculated as the average of Q values calculated for each single metabolite, whereas the Q value is the squared covariance between the metabolite and the outcome. Enrichment results reflect activation states of metabolic pathways among four experimental groups, based on curated pathway data from the Kyoto Encyclopedia of Genes and Genomes. Both intermediate and terminal metabolites were included in the pathway analyses.

Statistical Analysis

Data are presented as medians with interquartile ranges. Statistical analyses, excluding those related to metabolite profiling, were performed using EZR version 1.64 (https://www.jichi.ac.jp/saitama-sct/SaitamaHP.files/statmed.html). Intergroup comparisons were conducted using the U-test, whereas intragroup comparisons were assessed using the Wilcoxon signed-rank test or a mixed effects model. Statistical significance was defined as P < 0.05.

Results

Differentiation of ROs

The time course of RO differentiation is shown in Figure 1. ROs were generated using the serum-free floating culture of embryoid body–like aggregates with quick aggregation method (Figure 1A). During the first week of differentiation, dissociated cells aggregated to form embryoid bodies, which subsequently developed into ROs, consistent with previous findings.²¹ The outermost, clear layer of each organoid differentiated into retinal tissue across all iPS cell lines. Although minor variations in size and shape were observed among lines, retinal morphology remained consistent (Figure 1B). Immunostaining confirmed the presence of retinal marker–positive cells within the nuclear layer of the differentiated tissue (Figure 1C).

Figure 1.

Generation of retinal organoids (ROs) from induced pluripotent stem cells using the serum-free floating culture of embryoid body–like aggregates with quick aggregation (SFEBq) method. A: Schematic representation of the SFEBq protocol. B: Bright-field images of ROs at different stages of differentiation. Embryo body formation begins around day 7, and by day 20, a distinct outer neuroepithelial layer becomes visible, indicating retinal lineage specification. C: Immunohistochemical staining of the retinal markers retinal homeobox protein (Rx) (day 20), cone-rod homeobox protein (CRX), NeuN, and Math5 (all day 30). Scale bars: 500 μm (B); 50 μm (C). BMP4, bone morphogenetic protein 4; DMEM, Dulbecco’s modified Eagle’s medium; FBS, fetal bovine serum; FGFRi, fibroblast growth factor receptor inhibitor; GSK3i, glycogen synthase kinase 3 inhibitors; IMDM, Iscove's modified Dulbecco's medium; KSR, knockout serum replacement; RA, retinoic acid.

The morphology, TUNEL-positive cell ratio, and mtDNA copy number were compared between ROs derived from patients with LHON and healthy controls under varying concentrations and durations of SNP exposure. Figure 2 shows results from a single 24-hour exposure to increasing SNP concentrations. Figure 3 presents the results of three 24-hour exposures to 1 mmol/L SNP with 24-hour recovery intervals (Figure 3A). Figure 4 illustrates findings from a continuous 72-hour exposure to 1 mmol/L SNP (Figure 4A).

Figure 2.

Effects of acute sodium nitroprusside (SNP) exposure on retinal organoids (ROs). A: Bright-field images of ROs exposed to increasing SNP concentrations for 24 hours. Retinal layer clarity decreased at 1 mmol/L, with collapse of organoid structure at 5 and 10 mmol/L. B: Terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) staining of ROs showing dose-dependent increases in apoptosis, particularly in patient-derived organoids. C and D: Scatterplots of TUNEL-positive cell rates in whole organoids (C) and mitochondrial DNA copy number (D). Solid line: patient group; dotted line: control group. Scale bars: 500 μm (A); 200 μm (B).

Figure 3.

Intermittent sodium nitroprusside (SNP) exposure protocol: three 24-hour exposures to 1 mmol/L SNP with 24-hour recovery intervals. A: Schematic of repeated exposure to 1 mmol/L SNP. B: Bright-field images of retinal organoids (ROs) from each group. C: Representative terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) staining images in each group. D: Box-and-whisker plot of TUNEL positivity in control- and patient-derived ROs with and without SNP exposure. E: Box-and-whisker plot of mitochondrial DNA copy number with the same conditions as in D. ∗P < 0.05. Scale bars: 500 μm (B); 200 μm (C).

Figure 4.

Continuous sodium nitroprusside (SNP) exposure protocol: exposure to 1 mmol/L SNP for 72 hours without interruption. A: Schematic of continuous exposure to 1 mmol/L SNP. B: Bright-field images of retinal organoids (ROs) in each group. Arrows indicate retinal layer collapse in patient-derived organoids. C: Representative terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) staining images in each group. D: Box-and-whisker plot of TUNEL-positive cells in control- and patient-derived ROs with and without SNP exposure. E: Box-and-whisker plot of mitochondrial DNA copy number in the same conditions as in D. ∗P < 0.05. Scale bars: 500 μm (B); 200 μm (C).

Morphologic Changes in ROs following SNP Exposure

Across all lines, the outer retinal layer remained clear under baseline conditions (Figure 1B). After SNP treatment, this layer progressively darkened and thinned in a dose-dependent manner, indicating structural damage at SNP concentrations >1 mmol/L, especially at 5 and 10 mmol/L (Figure 2A). In ROs from the patient group subjected to three cycles of 24-hour exposure to 1 mmol/L SNP, the transparent layer was lost and organoids exhibited more pronounced atrophic changes (Figure 3B). Similarly, following 72-hour continuous SNP exposure, the loss of the clear outer layer was more evident in patient-derived organoids compared with controls (Figure 4B).

TUNEL-Positive Cells in SNP-Exposed ROs

TUNEL positivity increased with increasing SNP concentrations in the control and patient groups (Figure 2B). Median (interquartile range) TUNEL positivity rates at baseline were 0.460% (0.202% to 1.49%) in controls and 1.28% (1.01% to 2.50%) in patients, with no significant difference between the control and LHON groups at baseline (U-test, P = 0.077). However, the increase in TUNEL positivity with higher SNP exposure was significantly greater in the patient group (Figure 2C) (mixed effects model, P = 3.53 × 10⁻⁷).

Under the three-cycle exposure protocol, baseline TUNEL positivity (0 mmol/L SNP) did not differ significantly between groups (U-test, P = 0.196). Increased TUNEL-positive staining was observed in both groups after repeated SNP exposure (Figure 3C), with statistically significant increases in TUNEL positivity observed compared with the baseline (Figure 3D) (Wilcoxon signed-rank test control group; P = 3.71 × 10⁻², patient group; P = 3.91 × 10⁻²). Following 72-hour continuous SNP exposure, TUNEL-positive cells were primarily observed in patient-derived organoids (Figure 4C). A significant increase in TUNEL positivity was detected only in the patient group (Figure 4D) (Wilcoxon signed-rank test, control group; P = 0.734, patient group; P = 1.95 × 10⁻²).

mtDNA Copy Number in SNP-Exposed ROs

Following a single SNP exposure, the baseline mtDNA copy number (0 mmol/L) was significantly higher in patient-derived organoids compared with controls, with median (interquartile range) values of 52.1 (43.6 to 130.0) and 47.4 (43.1 to 58.1) copies/cDNA, respectively (U-test, P = 3.99 × 10⁻³). A similar pattern was observed after repeated 24-hour SNP exposures (Figure 3E) [36.3 (19.0 to 41.5) copies/cDNA in control ROs and 53.7 (48.7 to 61.0) copies/cDNA in patient ROs; U-test, P = 0.0244] and 72-hour exposure (Figure 4E) [47.4 (43.1 to 58.1) copies/cDNA in control ROs and 52.1 (43.6 to 130.0) copies/cDNA in patient ROs; U-test, P = 3.99 × 10⁻³]. Although mtDNA copy numbers increased with increasing SNP concentrations in both groups, the rate of increase did not differ significantly (mixed effects model, P = 0.162) (Figure 2D). In the three-cycle exposure protocol, mtDNA copy numbers increased significantly in both groups compared with baseline (Wilcoxon signed-rank test: control group, P = 3.91 × 10⁻³; patient group, P = 3.91 × 10⁻³) (Figure 3E). However, after 72-hour continuous exposure, a significant increase was observed only in the control group (Figure 4E) (Wilcoxon signed-rank test, control group; P = 1.25 × 10⁻², patient group; P = 0.301).

Metabolomic Analysis in SNP-Exposed ROs

Figures 5 and 6 present the results of metabolomic profiling. Figure 5 depicts a comparison of four experimental groups: control and LHON-derived ROs, each with and without 72-hour SNP exposure. Principal component analysis revealed clear group separation with no outliers, indicating distinct metabolic profiles (Figure 5A). Partial least squares discriminant analysis identified the 15 metabolites with the highest variable importance in projection scores, all exceeding the threshold of 1.0 (Figure 5B). Taurine had the highest variable importance in projection score, with its concentration decreasing significantly across the four groups in the following order: control without SNP exposure, control with SNP exposure, LHON without SNP exposure, and LHON with SNP exposure. Metabolites are listed in order of the largest concentration difference among groups, with taurine exhibiting the most pronounced change.

Figure 5.

Metabolomic changes in retinal organoids (ROs) following continuous 72-hour exposure to 1 mmol/L sodium nitroprusside (SNP). A: Principal component (PC) analysis of metabolite profiles. Red and green dots and ellipses represent control ROs without and with SNP exposure, respectively; blue and light blue dots and ellipses represent patient-derived ROs without and with SNP exposure, respectively. B: Top 15 metabolites identified via variable importance in projection (VIP) scores. Heat map squares indicate (from left to right) control organoids without SNP exposure (red), control organoids with SNP exposure (orange), patient-derived organoids without SNP exposure (blue), and patient-derived organoids with SNP exposure (purple). Taurine exhibits the highest VIP score. C: Hierarchical clustering of the top 15 differentially expressed metabolites across the four groups. SNP concentrations (0 or 1 mmol/L) are indicated above the group bars. GABA, γ-aminobutyric acid.

Figure 6.

Quantitative metabolite set enrichment analysis. Metabolite set enrichment was performed using the Small Molecule Pathway Database. Circles represent enrichment scores for metabolite sets, plotted in inverse order of P value across the four groups of retinal organoids [control organoids with and without sodium nitroprusside (SNP) exposure, and patient-derived organoids with and without SNP exposure].

Other key metabolites critical for RGC function and survival, including γ-aminobutyric acid, niacinamide (vitamin B3), and l-glutamic acid, also followed this declining trend. Hierarchical clustering analysis confirmed the top 15 metabolites with the most significant intergroup variations (Figure 5C). These analyses revealed the distinct retinal metabolite profiles among the four groups, highlighting consistent reductions in γ-aminobutyric acid, niacinamide, and taurine concentrations in the LHON groups. Additionally comprehensive metabolite profiling revealed significant differences in the metabolic signatures between LHON patient- and control-derived ROs, independent of SNP exposure (Figure 5, A and C). Figure 6 shows the results of metabolite set enrichment analysis based on organoid samples exposed to 1 mmol/L SNP for 72 hours. The most pronounced pathway alteration was observed in the Warburg effect, related to lactate metabolism. Additional significant differences were detected in pathways including bile acid biosynthesis, starch and sucrose metabolism, amino sugar metabolism, and taurine and hypotaurine metabolism, highlighting altered metabolic responses to oxidative stress in LHON-derived ROs.

Discussion

In this study, the effects of NO, delivered via SNP, on ROs derived from the iPS cells of patients with LHON were assessed. Results revealed that LHON-derived ROs were more susceptible to NO-induced damage compared with control ROs, exhibiting altered metabolic profiles, particularly in metabolites associated with neuronal homeostasis. Although only one case report has previously suggested a potential link between LHON and NO, it relied on peripheral blood rather than retinal tissue for analysis.²² Thus, the current study represented a novel contribution to understanding LHON pathophysiology using patient-derived retinal tissue.

The rationale for investigating NO originated from a key feature of LHON: the presence of parapapillary telangiectasia during the acute phase. These dilated capillaries around the optic disc appear even before the onset of visual loss. Notably, they do not leak fluorescein dye during angiography, suggesting intact blood-retinal barrier function, unlike other optic nerve disorders, including optic neuritis and anterior ischemic optic neuropathy, or even papilledema due to elevated intracranial pressure. Given NO's dual role as a vasodilator and a neuroprotective agent at low concentrations, with neurotoxicity occurring at elevated levels,¹⁰ it was hypothesized that NO overproduction dilates capillaries around the optic nerve head and exerts deleterious effects on RGCs in patients with LHON. The present findings support this hypothesis, particularly the latter notion, demonstrating that NO disrupts retinal metabolism and compromises RGC survival in patients with LHON harboring the m.11778G>A mutation. A key insight from this study is that SNP-induced damage extends beyond RGCs to the entire retinal tissue, and that intrinsic metabolic dynamics differ between patient- and control-derived ROs. Although prior studies have focused primarily on RGCs, clinical evidence indicates the broader involvement of retinal or extraretinal cells in LHON. For instance, in typical LHON cases, only the RGCs in the papillomacular bundle degenerate at disease onset, likely due to the structural fragility of their unmyelinated nerve fibers.²³ This suggests that although RGCs are the primary site of injury, the underlying vulnerability may reflect systemic or cell-intrinsic abnormalities rather than isolated RGC pathology. Therefore, this study focused on whole ROs to more comprehensively capture disease-related changes.

Importantly, no differences in morphology or retinal differentiation marker expression were observed between LHON and control ROs under standard culture conditions (Figure 1, Figure 2, Figure 3, Figure 4). This aligns with the pathogenesis of LHON, as individuals with LHON mutations typically show no retinal abnormalities before disease onset, except for parapapillary telangiectasia, despite the genetic mutation being present. This feature distinguishes LHON from other inherited retinal diseases. Nevertheless, even under physiological conditions, significant differences in mtDNA copy number and metabolite profiles were observed between LHON and control ROs, consistent with previous studies involving peripheral blood cells,^24,25 suggesting that functional abnormalities exist in LHON-derived tissues without apparent structural or morphologic defects.

Various genetic mutations contribute to the development of LHON. In total, >90% of the cases have one of the three major mtDNA mutations: m.3460 G>A, m.11778G>A, or m.14484 T>C, whereas the remaining 10% of the cases have other rare mtDNA or autosomal recessive nuclear DNA mutations.²⁶ These mutations impair electron transfer complex 1 activity, disrupting electron transfer and ATP production in the mitochondria, elevating reactive oxygen species production, and ultimately triggering caspase-independent apoptosis in RGCs.²⁷ However, prior epidemiologic studies suggest that the prevalence of mtDNA mutations may be much higher than previously estimated (up to 1 in 800 individuals),^28,29 implying low disease penetrance and the need for additional genetic, epigenetic, or environmental factors to manifest the LHON phenotype. Known triggers include heavy smoking and alcohol consumption,³⁰ antibiotic use or trauma,^31,32 and sex hormone influences.^5,33

Consistent with prior findings in peripheral blood,^24,25 the present results showed that mtDNA copy number was significantly higher in LHON-derived ROs compared with controls under baseline conditions. This may reflect a compensatory mechanism that preserves mitochondrial function by increasing the mtDNA copy number. Interestingly, although lower SNP concentrations induced cell death in patient-derived organoids, the rate of mtDNA copy number increase in response to SNP was comparable between the control and patient groups. This aligns with a previous study reporting reduced viability of LHON-derived lymphoblasts after NO exposure.²² Furthermore, repeated intermittent SNP exposure (1 mmol/L for 72 hours) increased cell death and mtDNA copy number in both groups, whereas continuous SNP exposure (1 mmol/L for 72 hours) increased mtDNA copy number only in the control group, with a significant increase in cell death only observed in the patient group. Collectively, these results highlight three key points: i) at baseline, LHON-derived ROs exhibit elevated mtDNA copy numbers, potentially as a compensatory response to mitochondrial dysfunction; ii) these ROs are more vulnerable to NO-induced stress compared with control ROs; and iii) the ability to sustain compensatory mtDNA up-regulation under prolonged NO exposure is impaired in LHON-derived ROs.

In the present study, a potential mechanism linking apoptosis and mtDNA copy number is proposed in the context of LHON. mtDNA copy number is highest in LHON carriers, followed by healthy individuals, and lowest in patients with LHON, suggesting that energy production is compensated by increasing mtDNA copies.^32,33 These findings indicate that maintaining cellular energy production by increasing the mtDNA copy number benefits long-term survival.

In clinical practice, fundus findings support the notion that NO is released before the onset of visual symptoms in LHON. In typical LHON cases, visual dysfunction first appears in the lateral eye, followed by the contralateral eye after several weeks or months. However, even before symptom onset in the contralateral eye, signs of parapapillary telangiectasia, specifically in the papillomacular bundle, are already present. Therefore, NO is released before symptoms occur. Moreover, imaging studies have shown that inner retinal thinning, indicative of RGC death, is evident in both eyes at the time of onset.

The current findings suggest that prolonged NO exposure, coupled with a failure to further increase mtDNA copy numbers, may impair the energy compensation system in LHON-derived retinal tissue, potentially contributing to the acute onset of the disease. In the present metabolomic analysis, key metabolites critical for retinal function and survival, such as taurine, γ-aminobutyric acid, niacinamide, and glutamate, were significantly reduced in patient-derived organoids, particularly after prolonged NO stress. Initially, it was hypothesized that SNP exposure alters the metabolic pathways of ROs; however, the baseline metabolic signatures of LHON and control ROs were found to differ (Figure 5). Notably, taurine levels were lower in LHON ROs than in control ROs before SNP exposure, and this difference increased with SNP treatment.

This altered metabolic profile likely reflects intrinsic differences in how patient-derived organoids respond to SNP exposure. Taurine and niacinamide, or nicotinamide, levels are known to be reduced in the peripheral blood of patients with LHON.³⁴ Taurine is neuroprotective in the retina, partly by binding with bile acids to form tauroursodeoxycholic acid, an endogenously biosynthesized compound.³⁵ Nicotinamide has demonstrated neuroprotective effects in glaucoma models³⁶ and shown promise in human studies.³⁷ Thus, the reduced concentrations of these metabolites in patient-derived ROs, particularly under prolonged NO exposure, suggest that impaired retinal metabolism is exacerbated by both extrinsic and intrinsic factors.

Although patient-derived ROs showed a distinct metabolite profile at baseline, there was no significant difference in apoptosis rates compared with controls under physiological conditions. However, exposure to SNP significantly increased cell death in patient-derived ROs, consistent with the clinical scenario in which LHON phenotypes are typically triggered by environmental or physiological stressors. Importantly, the current study involved targeting the entire retinal tissue, not just RGCs. Thus, the findings indicate that mitochondrial dysfunction in LHON affects not only RGCs but also other retinal cell types, contributing to a broader, subclinical impairment of cellular metabolism, including taurine metabolism. Indeed, some clinical studies have indicated that photoreceptor cells may also be compromised in patients with LHON.^9,38

This study has several limitations. First, from a clinical point of view, this study did not address whether the retinal capillaries of patients with LHON produce or overproduce NO. Second, although iPS cell–derived ROs are a valuable tool to study the pathogenesis of LHON, they do not fully replicate the complexity of the human retina.³⁹ Furthermore, as the organoids used in this study were at an early stage of differentiation,³⁹ validation of the results with more mature organoids is required. Finally, RGCs' responses to NO stress were not specifically assessed. Although studies are limited, one report⁴⁰ has shown that NO synthase inhibition reduces RGC survival in a dose-dependent manner in rat retinal explants, supporting the dual role of NO in retinal neurobiology. Nevertheless, by evaluating whole-retina organoids, the present study provides a novel perspective on understanding the broader pathophysiological mechanisms of LHON.

In conclusion, ROs derived from the iPS cells of patients with LHON showed increased vulnerability to SNP-induced NO stress, evidenced by changes in cell viability and mtDNA copy numbers, as well as reduced levels of key neuroprotective metabolites. These findings highlight the critical role of NO and impaired metabolic adaptation in the pathogenesis of LHON. Further investigation of the impact of NO could be a breakthrough in the treatment of LHON.

Disclosure Statement

None declared.

Supplemental Data

Supplemental Figure S1.

Characterization of induced pluripotent stem (iPS) cells derived from the patient group (patients 1 to 3) and control 1. A: iPS cell colonies of each line. B: Immunohistochemistry with pluripotency markers OCT3/4, NANOG, TRA-1-60, and SSEA4, indicating the undifferentiated status of the iPS cell colonies. C: Immunohistochemistry with alpha-fetoprotein (AFP), smooth muscle actin (SMA), and neuron-specific class III beta-tubulin (Tuj-1), indicating the pluripotency of the iPS cell colonies. Scale bars: 500 μm (A); 100 μm (B); 50 μm (C).

Supplemental Figure S2.

A representative terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) staining image of a retinal organoid. The outermost layer of the retinal organoid (arrowheads) was used for TUNEL analysis. Green, TUNEL positive; blue, DAPI.