Idebenone Protects Photoreceptors Impaired by Oxidative Phosphorylation Disorder in Retinal Detachment

Abstract

Purpose

Oxidative phosphorylation (OXPHOS) is an aerobic metabolic mechanism, and its dysfunction plays an important role in the pathological changes of ischemic diseases. However, systematic studies on the occurrence of retinal detachment (RD) are lacking.

Methods

Single-cell RNA sequencing (scRNA-seq) of the human retina was performed to detect the metabolic changes of various retinal cells after RD. In this study, animal experiments were conducted to explore the OXPHOS activity after RD. In addition, idebenone, a coenzyme Q10 (CoQ10) analog currently used to treat Leber hereditary optic neuropathy (LHON), was used to improve the OXPHOS disorder in experimental RD model.

Results

ScRNA-seq revealed abnormal energy metabolism and OXPHOS pathways in retinal cells after RD. Adenosine triphosphate (ATP) and reactive oxygen species (ROS) are the main products of OXPHOS, the mouse RD model indicated that the rise in ROS levels may have a greater impact on photoreceptors in the early stage, whereas decreased ATP synthesis was observed in the later stage; these changes threaten the function and morphology of the retina. Idebenone was administered to model mice intragastrically, leading to reduced ROS levels in the early stage post-RD and improved ATP synthesis in the later stage, which was closely related to the maintenance of mitochondrial morphology.

Conclusions

OXPHOS disorder leads to photoreceptor degeneration after RD, which can be alleviated by improving OXPHOS function.

Retinal detachment is a serious blinding eye disease with a rising incidence of about one to two per 10,000 people.^1,2 The main mechanism underlying visual impairment in RD is the cell death of cones and rods.³ RD leads to separation between the retinal nerve epithelium and pigment epithelium, resulting in ischemia and hypoxia of the outer retinal layers, thereby causing photoreceptor damage. Previous studies have revealed increased ROS levels in RD, which damage photoreceptors.⁴ The abnormal rise of ROS levels may be attributed to multiple factors, the most important of which is the complex I of the electron transfer chain (ETC) of OXPHOS located in the inner mitochondrial membrane.⁵

Mitochondria are not only the main source of ROS but are also responsible for the synthesis of ATP. In mitochondria, OXPHOS is a critical process in aerobic ATP synthesis⁶; therefore the disruption of mitochondrial homeostasis decreases energy and causes cell damage.⁷ The high energy requirements of retinal tissues are met by both aerobic and anaerobic metabolism.⁸ Under ischemic and hypoxic conditions, the aerobic synthesis of ATP is reduced,⁹ which inevitably leads to the decline of retinal function.

Idebenone (IDB), a coenzyme Q10 analog, is also an important reducing agent.¹⁰ Idebenone has been applied in various models of ischemia and hypoxia to relieve the tissue and cell damage caused by oxidative stress.^11,12 Its protective effects are mainly mediated by improving the electron transfer efficiency in the process of OXPHOS. This treatment scheme has been clinically applied in the treatment of LHON.¹³

In this study, scRNA-seq revealed that the differential genes of various retinal cells, such as cones and rods, were enriched in energy metabolism, ATP synthesis, and mitochondrial complex I regeneration pathways after RD. These findings suggest that the mechanism of photoreceptor damage caused by RD is most likely related to abnormal OXPHOS, which has been confirmed in animal models. Moreover, idebenone has been applied to alleviate OXPHOS dysfunction, indicating a potentially new therapeutic option for preserving retinal function in RD.

Methods

Clinical Information

Retinal tissue collection and research were approved (no. 2023KY340) by the Hospital Ethics Review Committee of the First Affiliated Hospital of the University of Science and Technology of China (USTC), and all patients and donor families provided informed consent. In the case of rhegmatogenous retinal detachment (RRD), only the aperture cap and the crimped retina located in the peripheral retinal tears were obtained during vitrectomy (the parts that would have been removed during surgery due to the inability to reattach). The donor was obtained peripheral retina (similar location to RRD patients). The tissues of three RRD patients (Table 1) were mixed because the individual tissues were too small; the peripheral retina of one donor (Table 2) with no history of eye disease was obtained. All tissues obtained were placed in a preservation solution and stored at −80°C until detection. All experiments were conducted according to the ethical system of the First Affiliated Hospital of the USTC.

Table 1.

Information of RRD Patients

Group	Age	Time Post-RD	Gender
RRD1	17 y	1 week	Male
RRD2	50 y	1 week	Male
RRD3	59 y	1 month	Male

Table 2.

Information of the Donor

Group	Age	Gender	Time From Death to Tissue Collection	Death Reason
Donor	54 y	Male	1 hour	Stroke

Animal Ethics

All experiments involving animals were approved by the Animal Research Institute Committee of the First Affiliated Hospital of USTC and strictly adhered to the Association for Research in Vision and Ophthalmology statement. The experimental animals were eight to 10 weeks old, healthy, male C57BL/6J mice (weighing 20–25 g), which were provided by the First Affiliated Hospital of USTC. The animals were allowed free access to water and food and were housed in a specific pathogen-free mouse facility under a 12-hour light /12-hour dark cycle.¹⁴

Single-Cell RNA Sequencing

Single-cell RNA sequencing was performed according to the manufacturer's instructions. All retinal tissues were digested and processed into cell suspension, and cell viability was measured to ensure that the proportion of active cells was greater than 90%. Single-cell gel beads were generated in solution according to the manufacturer's protocol. The library was sequenced using the Illumina Novaseq6000 sequencer. The sequencing depth was at least 10,0000 reads per cell, with pair-end 150 bp (PE150, performed by Beijing CapitalBio Technology, Beijing, China). A cell ranger counting module was used for comparison.¹⁵ Subsequently, the characteristic bar code matrix was generated and clustering was determined. Uniform manifold approximation and projection was used for visualization.¹⁴

RD Model

Mice were anesthetized with pentobarbital (50 mg/kg), and the pupils were dilated with 0.5% tropicamide. The eyelashes were cut, iodophor was used to disinfect the eyes, and levofloxacin eye drops were used to rinse the conjunctival sac (Suzhou Santeng Pharmaceutical Co., LTD., Jiangsu, China). Sclerotomy was made 1 mm behind the corneal limbus with a 30G needle, and approximately 4 µL of 10 mg/mL sodium hyaluronate (Albomed, Schwarzenbruck, Germany) was injected slowly and gently into the subretinal space, thereby completely detaching the sensory retina from the RPE.¹⁶ RD models were all established in the right eyes (Supplementary Fig. S1), whereas the left eyes underwent all procedures except for subretinal injection were used as control.^17–19 Mice with cataracts, retinal hemorrhage, corneal abnormalities, and endophthalmitis were excluded.²⁰

Tunel Staining

Tunel staining was carried out to observe the apoptosis levels of outer retinal cells in each group. After fixation, the collected eyeballs were sliced into paraffin sections and treated at 37°C for one hour according to the manufacturer's instructions. Then, the sections were stained with DAPI and observed under a fluorescence microscope after closure, and the data were processed by image J software.²¹

Western Blot

After establishing the mouse retinal detachment model, retinal tissues were obtained following the experimental procedure, and retinal proteins were extracted by radioimmunoprecipitation assay buffer after ultrasonic homogenization. Electrophoresis was performed at constant voltage and the gel was removed from the instrument. The proteins were then transferred to a blocking membrane of 0.22um under constant current. After transfer, the samples were washed three times with Tris buffered brine Tween (TBST) for 10 minutes each, and sealed with 5% skim milk for one hour. Then, the samples were washed three times again with TBST, and the membrane was incubated with NDUFB8 (1:1000; Proteintech, Wuhan, Hubei, China), SDHB (1:1000; Proteintech), UQCRFS1 (1:1000; Proteintech), COX4 (1:1000; Proteintech), ATP5A (1:1000; Proteintech), and β-actin (1:1000; Proteintech) antibodies and were incubated at 4°C overnight. The primary antibodies were removed, washed with TBST three times, and incubated with the secondary antibodies labeled with hrp (1:5000; Proteintech) at room temperature for one hour. The protein is revealed by enhanced chemiluminescence.²²

Oxygen Consumption Rate (OCR)

The mice were acclimated to darkness for 12 hours overnight before the experiment. The mice were anesthetized shortly after the light was turned on, and the collected retina tissues were preserved on ice. Two 1 mm–diameter tissue samples were collected around the optic nerve of each retina using a 1mm biopsy punch. Each tissue was transferred into a 24-well islet capture microplate. Each pore contained 500 µL intracellular fluid (135 mmol/L NaCl, 10 mmol/L HEPES, 3.1 mmol/L KCl, 0.5 mmol/L KH₂PO₄, 1.5 mmol/L CaCl₂, 1.2 mmol/L MgSO₄, 6 mmol/L glucose). NaOH was used to regulate the pH to 7.4. After transferring the tissue, the islet capture microplate was incubated at 37°C for one hour. The OCR was measured by Seahorse extracellular flux Analyzer XF24 (Agilent Technology, Tokyo, Japan) and the Mito Stress Test Kit (101122-100; Agilent Technology). After baseline OCR was determined, 20 µmol/L oligomycin was added to each well to inhibit ATP production, then 5 µmol/L carbocyano4 –(trifluoromethoxy) phenylhydrazine (FCCP) was added, and finally 2 µmol/L rotenone and 2 µmol/L antimycin A (R + A). Five cycles of measurement were performed after each treatment, with each data point representing the average of six single wells per processing. Normalized results were calculated based on protein levels.^14,23,24

ATP Assay

After successfully collecting retinal samples from each group of mice, tissue homogenate was prepared according to the manufacturer's instructions (Solarbio Technology, Beijing, China), and ATP was extracted. The absorbance value of the ATP extract from each well was detected by a microplate reader, and the ATP content of each sample was calculated according to the reference formula provided by the manufacturer.²⁵

ROS Aassay

The fluorescence probe provided in the kit (Bestbio, Shanghai, China) was added to the prepared tissue homogenate and the fluorescence intensity of each hole was measured using the microplate reader. Finally, the ROS content of each sample was calculated based on protein levels.⁴

Transmission Electron Microscope (TEM) Assay

The morphological changes of mitochondria in retinal photoreceptors were observed by TEM. Briefly, eyeballs were removed immediately after mice were killed and fixed with 4% glutaraldehyde for 24 h. The retinas were collected, fixed with 1% osmium tetroxide for one hour, dehydrated with gradient ethanol, infiltrated with propylene oxide and epoxy resin, and embedded in an oven at 40°C and 60°C for 12 and 24 hours, respectively. The embedded tissue was removed and trimmed before sectioning. The sections were recovered with copper mesh, stained, observed, and imaged by TEM.²²

Statistics

The count data were expressed as the mean ± SD. The t-test was used for comparison between two groups, whereas the analysis of variance (ANOVA) was used for comparison between multiple groups. Quantitative plots were analyzed using GraphPad Prism9. In this study, a P value of 0.05 was considered statistically significant.

Results

Abnormal Energy Metabolism of Various Retinal Cells After RD

The mixed samples from three RRD patients and the retinal tissue from one donor were subjected to scRNA-seq. The results revealed that the retinal tissue samples of RRD patients had a total of 6930 cells, and the donor retinal tissue sample had a total of 12,554 cells. After filtering, all cells were divided into eight clusters (Supplementary Fig. S2A). A total of 268 up-regulated genes and 166 down-regulated genes were found after RD (Supplementary Fig. S2B). The differential biological process (BP) and cellular component (CC) pathways were primarily involved in the ATP-producing and oxidoreductase complex (Supplementary Figs. S2C, S2D). The abnormal energy metabolism and OXPHOS of rods, bipolar cells, and Müller cells in RD were observed (Supplementary Fig. S3).

The OCR of the Retinal Tissues of Mice Decreased After RD

To further explore the changes in OXPHOS after RD, an OCR assay was performed to detect the OCR of mouse retinal tissue. The results indicated that the baseline of the attached group was significantly higher than in the retinal tissues at day 3 post-RD (Figs. 1A, 1B) in OCR. In addition, the OCR curves of each group were plotted after oligomycin, FCCP, and R+A were added, respectively. The maximal respiration in the attached group was also higher than that on day 3 post-RD (Figs. 1A, 1D). These results suggested impaired retinal OXPHOS function after RD.

Figure 1.

The OCR of the retinal tissues of mice decreased after RD. (A) OCR time series curves of the attached group (n = 6) and three days post-RD (n = 6). (B) Baseline OCR at day 3 (55.11 ± 23.93, P = 0.031) post-RD was significantly lower than that of the attached group (102.32 ± 37.75). (C) ATP production at three days (8.82 ± 5.10) post-RD was not significantly different from that of attached group (9.79 ± 7.96). (D) The maximal respiration at three days (46.29 ± 20.55, P = 0.018) post-RD was significantly lower than that of the attached group (92.53 ± 32.63).

Abnormal Mitochondrial Morphology and Function Were Observed in Photoreceptors in the Experimental RD Model

Because ATP and ROS are both products of OXPHOS, impairment of OXPHOS results in decreased ATP production and abnormally raised ROS levels.^26,27 Therefore this study examined the content of ATP and ROS, and the time series curves were drawn to evaluate the function of OXPHOS (Fig. 2A). The results revealed that ROS increased significantly on the third and fifth day after RD. ATP production showed no significant change on day 1, day 3, and day 5, but it was significantly lower than that of the attached group on day 7 post-RD (Fig. 2A).

Figure 2.

Abnormal mitochondrial morphology and function were observed in photoreceptors in the experimental RD model. (A) Time series curves of ROS and ATP post-RD: ROS increased significantly at day 3 (1.67 ± 0.28, P = 0.021, n = 4) and day 5 (1.98 ± 0.39, P = 0.007, n = 4) post-RD compared with the attached group. ATP had no significant change on day 1 (0.87 ± 0.12, n = 4), day 3 (0.81 ± 0.14, n = 4) and day 5 (0.74 ± 0.09, n = 4) post-RD, and was significantly lower than that of attached group on day 7 (0.59 ± 0.10, P = 0.04, n = 4) post-RD. (B–D) Morphological changes of mitochondria post-RD: on the third (3.44 ± 0.35, P = 0.002, n = 3) and seventh (2.52 ± 0.47, P = 0.013, n = 3) day post-RD, the mitochondria became vacuolar and the proportion of irregular mitochondria increased significantly compared with the attached group (n = 3). ^⋀ indicates abnormal mitochondria; * indicates normal mitochondria.

Considering that OXPHOS is carried out in mitochondria, the mitochondrial morphology of the photoreceptors was observed by TEM. The results revealed that the mitochondria of photoreceptors were significantly vacuolated on days 3 and 7 after RD (Figs. 2B–D). These results suggested that post-RD OXPHOS dysfunction may induce changes in ROS and ATP production, which may also be associated with mitochondria-related photoreceptor damage.

OXPHOS Dysfunction in the Experimental RD Model

To verify the functional changes in retinal OXPHOS in RD, a mouse RD model was established and the activity of OXPHOS complex I–V was detected. Western blot tests were performed on mouse retinas, and NDUFB8, SDHB, UQCRFS1, COX4, and ATP5A subunits were used to detect the protein expression of complex I-V, respectively. The results revealed no significant change in the expression of OXPHOS complex I-V in the retina of mice on the first, third day after RD (Figs. 3A–F). In contrast, complex I showed significantly decreased levels on the seventh day post-RD while complex II reduced from day 5 post-RD (Figs. 3A–F), although the expression levels of complex III-V remained unchanged. These results suggested dysfunctional retinal OXPHOS in RD.

Figure 3.

OXPHOS dysfunction in the experimental RD model. (A–F) Changes of OXPHOS complex activity after RD (n = 3): The activities of NDUFB8 (complex I) decreased significantly on the seventh day (0.54 ± 0.14, P = 0.007) post RD, and SDHB (complex II) decreased significantly on the fifth day (0.46 ± 0.14, P = 0.003) and seventh day (0.37 ± 0.12, P < 0.001) post RD; UQCRFS1 (complex III), COX4 (complex IV), and ATP5A (complex V) had no significant changes compared with control group in whole time series.

IDB Protects OXPHOS After RD

IDB is currently the only drug approved in Europe for the treatment of LHON,²⁸ which can improve the function of ganglion cells in patients suffering from complex I deficiency. Therefore IDB was used to treat experimental RD model mice in our research. The results indicated that IDB significantly reduced retinal ROS levels on day 3 and improved ATP decline on day 7 after RD (Figs. 4A, 4B). These findings indicated that IDB could mitigate retinal damage caused by impaired OXPHOS after RD.

Figure 4.

Idebenone preserved OXPHOS function post-RD. (A) IDB mitigated the rise of ROS at day 3 post-RD: the ROS levels in the IDB group (1.10 ± 0.24, n = 3) were significantly lower than that in the RD group (1.52 ± 0.13, P = 0.036, n = 3). (B) IDB mitigated ATP decline on day 7 after RD: ATP content was significantly higher in the IDB group (0.75 ± 0.07, n = 4) compared to the RD group (0.54 ± 0.09, P = 0.045, n = 4). (C) IDB reduced photoreceptor apoptosis on day 3 post-RD: the apoptotic signal was significantly reduced in the IDB group(12.24 ± 3.10, n = 3) compared with the RD group (34.35 ± 7.32, P = 0.001, n = 3).

Apoptosis is considered one of the main injury modes of retinal photoreceptors in RD.²⁹ Therefore TUNEL staining was performed on mouse retina samples to investigate the protective effect of IDB against photoreceptor damage. The apoptotic signals in the retinal photoreceptors of mice three days post-RD were significantly increased compared to the attached group, while IDB significantly reduced photoreceptor apoptosis (Figs. 4C, 4D). These results suggested that IDB alleviates photoreceptor damage induced by RD.

Discussion

This study demonstrated through human tissue tests and animal experiments that RD-induced OXPHOS dysfunction may lead to increased retinal ROS, decreased ATP synthesis, and mitochondrial damage in photoreceptors. Interestingly, the rise of retinal ROS was found to precede the decline of ATP synthesis in animal models of RD and were not synchronized after RD occurred. Hence, early RD photoreceptor damage caused by impaired OXPHOS may be primarily attributed to oxidative stress rather than energy deficiency. However, the application of idebenone effectively improved the rise in ROS and the ATP decrease and saved photoreceptors. These results indicated that OXPHOS dysfunction post-RD was closely related to the mechanism of photoreceptor injury.

Recently, scRNA-seq was used to investigate the pathogenesis of many diseases, which overcomes the data bias caused by tissue heterogeneity.³⁰ In this study, retinal tissues from 3 RRD patients and a donor without eye disease were used for scRNA-seq. The results revealed differential gene enrichment in various cells, including photoreceptors and microglia, focusing on energy metabolism and the OXPHOS pathway (Supplementary Fig. S3). Photoreceptors are metabolically active, located in the outer layer of the retina, and are irrigated by the choroidal blood supply system.³¹ Therefore photoreceptors are susceptible to hypoxia, resulting in an inevitable decline in aerobic metabolic activity after RD. Although the specific activation mechanism remains to be explored, OXPHOS disorders have been found to participate in the activation of microglia in cerebral ischemic diseases, and microglia are involved in the neuroinflammation of photoreceptors after RD.³² Therefore the remodeling effects of impaired OXPHOS on retinal microglia is traceable, but the underlying mechanism requires further research.

OCR is a direct reflection of the OXPHOS function.³³ In this study, the baseline and maximum value of OCR in the attached retina were significantly higher than RD (Fig. 1A), which further verified the retinal OXPHOS dysfunction after RD. However, the maximum respiratory value measured after the addition of FCCP was not significantly higher than the baseline, which may be related to the high concentration of oligomycin. These findings were also in agreement with the results of previous studies.^14,33

ROS and ATP are important products of OXPHOS, both of which are essential for cell metabolism. ROS is an oxidizing compound that plays a crucial role in oxidative stress and is also involved in a variety of cellular damage mechanisms, including apoptosis, oxidative stress, and ferroptosis.^34–37 Under physiological conditions, ROS acts as an active oxidized substance and participates in many intracellular reaction processes. However, the superphysiological concentration of ROS not only directly reacts nonspecifically with proteins, nucleic acids, lipids, and carbohydrates, but also promotes the generation of other toxic substances.³⁸ Hypoxia impaired OXPHOS results in abnormally elevated ROS levels and may cause mitochondrial and cellular damage.³⁹ In the retina of mice, ROS was abnormally elevated from day 3 post-RD (Fig. 2A), which was consistent with the results of previous studies.¹⁸ Nonetheless, ATP did not decrease significantly at this time, which seemed to indicate that photoreceptor damage mediated by impaired OXPHOS was mainly attributed to the abnormal increase in ROS, rather than the lack of energy supply. Still, ATP showed a significant decrease on the seventh day post-RD, and retinal ROS levels reduced from peak at this time, which might be related to the decreased expression of complex I and II (Figs. 3A–C).

OXPHOS is carried out in mitochondria. Abnormal mitochondrial function and morphology result in impaired OXPHOS, which in turn impairs mitochondria.⁴⁰ Therefore mitochondrial morphology reflects cell activity and OXPHOS function. In this study, the mitochondrial morphology of mouse retina showed vacuoles from the third day post-RD, accompanied by the disappearance of the mitochondrial ridge (Fig. 2B). These changes indicated impaired mitochondria after RD, which is similar to the results of previous studies,⁴¹ what difference is that the research found that COX4 decreased three days after RD in experimental rat model, whereas in our study, COX4 did not decrease significantly within seven days after RD in mouse retina, which may be related to the different model animals and experimental methods we selected.

OXPHOS disorder is an important cause of energy metabolic reprogramming and is closely related to the pathogenesis of many diseases, including tumors and neurodegeneration.^42,43 In this research, mice were used to explore the function of retinal OXPHOS after RD. The results revealed that the expression of OXPHOS complex I and complex II decreased after RD model (Figs. 3A–C). As the anchor site of the OXPHOS complex, the dysfunction of the mitochondria may cause the decline of OXPHOS complexes as a result of hypoxia.⁴⁴ Complex I has been identified as an essential component of OXPHOS, a major source of mitochondrial ROS, and is involved in approximately 40% of the ATP produced by OXPHOS.²⁷ Thus complex I deficiency directly affects the production of ROS and ATP, which were both confirmed in the time generation curve after RD (Fig. 2A). As opposed to ROS, the total amount of retinal ATP did not decrease significantly in the early RD stage, whereas cones and rods showed obvious apoptosis at this time (Fig. 4C). The results indicated that photoreceptor damage mediated by OXPHOS disorder in early RD might not be due to insufficient energy synthesis, but caused by excessive ROS increase.

LHON refers to a visual dysfunction caused by an inherited mitochondrial complex I defect. Notably, idebenone has shown efficacy in treating the disease, significantly improving the function and morphology of retinal ganglion cells.^45,46 Its role may be related not only to its antioxidant properties but also to its irreplaceable electron transport mechanism in OXPHOS.⁴⁷ In this study, idebenone was used to treat RD model mice, and the results revealed significantly reduced apoptosis of photoreceptors in the outer layer of the retina. Moreover, the OXPHOS-induced ROS rise on day 3 and ATP decline on day 7 of RD were alleviated. These results are consistent with those found in central nervous system diseases and LHON, thereby indicating the potential of idebenone in the treatment of RD.

Nevertheless, the limitations of the current study should be acknowledged. For example, the sample size for scRNA-seq is small at one RD (mixed three donors into one) to one control, and control sample for scRNA-seq was from a deceased donor, although the tissue was obtained within one hour from donor's death, the postmortem effects on retinal OXPHOS cannot be completely ruled out. The time series curves of total ROS and ATP content in the retina of mice after RD have been determined, the specific mechanisms underlying these changes remain unclear. In addition, the specific mechanism of photoreceptor damage caused by OXPHOS disorder after RD is unelucidated. Moreover, the maximum value of OCR in each group was not significantly higher than the baseline, which may be attributed to the high concentration of oligomycin. Overall, photoreceptor damage caused by OXPHOS disorder after RD requires further research to determine the mechanism. These findings will lay a foundation for the clinical application of idebenone in RD.