This study was approved by the Animal Ethics Committee of Central South University (NO. 2019sydw0195). All experiments were conducted in accordance with the criteria of the National Institutes of Health Guide for the Care and Use of Laboratory Animals. Male C57BL/6 mice aged 6–8 weeks (weight, 18–22 g) were purchased from Hunan SJA Laboratory Animal Co., Ltd. (Changsha, China). Mice were maintained in individual ventilated Plexiglas cages under a 12 h light/dark cycle at room temperature in the Department of Laboratory Animals of Central South University throughout the experiment.

NaIO₃ (40 mg/kg) (Sigma-Aldrich, St. Louis, MO, USA) and EC (100 mg/kg/day) (Sigma-Aldrich, St. Louis, MO, USA) were dissolved in sterile normal saline and sterile drinking water, respectively. The concentrations of drugs were chosen according to the body of literature (19, 20).

To evaluate the safety of EC, mice were randomly allocated into one of two groups: (1) the control group, which received only sterile drinking water by oral gavage, and (2) the EC group (at a dose of 100 mg/kg/day, by oral gavage). To assess the efficacy of EC, mice in the control group were given tail vein injections containing only 10 mL/kg normal saline; mice in the NaIO₃ or EC group were treated with drinking water or 100 mg/kg/day EC orally via gastric gavage respectively, and given 40 mg/kg NaIO₃ by tail vein injection half an hour after gastric gavage. Gavage administration of water or EC was given once daily for 7 days, and subsequent experiments were then carried out.

Mice were sacrificed and transcardially perfused with 10 mL pre-cooled normal saline. The eyes were enucleated and fixed in FAS eye fixation fluid (Servicebio, Wuhan, China) for 24 h. Next, they were gradient dehydrated and embedded in paraffin. The paraffin eyes were sliced into 3 μm sections on the sagittal plane through the optic nerve (Leica, Wetzlar, Germany). Hematoxylin and eosin (HE) staining was carried out using an HE staining kit (Servicebio, Wuhan, China).

Light microscopy (Leica, Wetzlar, Germany) was used to obtain retinal images located 200–300 μm from the optic nerve head, which were compared for morphological observation and nuclei counting in the outer nuclear layer (ONL). Using Image Pro Plus 6.0, four images were grabbed at ± 250 and ± 500 μm per eye for quantification of the total area of drusen-like deposits. Also analyzed were the average thickness and the thickness at ± 150, ± 300, ± 450, ± 600, ± 750, ± 900, ± 1,050, ± 1,200, ± 1,350, ± 1,500 μm from the optic nerve of the ONL. The IS/OS were also analyzed.

After 7 days of oral gavage of water or EC, mice were anesthetized by an intraperitoneal injection of 0.2 ml/ 20 g of 1.0% pentobarbital sodium (Merck, Darmstadt, Germany). Pupils were dilated with one drop of 0.5% compound tropicamide (Santen, Tokyo, Japan) 5 mins before examination. Hydroxypropyl methylcellulose ophthalmic demulcent solution was applied to the corneal surface to form a uniform optical interface. Ocular fundus and OCT images were captured using a Micron IV retinal imaging system and Reveal OCT2 imaging system (Phoenix Research Labs, Pleasanton, CA, USA), respectively.

To measure the percentage of retinal degeneration area, color fundus images were converted to gray scale 16-bit images and threshold using Image Pro Plus 6.0. Region of degeneration was then determined. The area of optic disc was substracted, and ratio of degeneration area/total area was calculated.

Following 8–12 h pre-adaptation to darkness, the mice were anesthetized, and the pupils were maximally dilated under red light. Full-field ERGs were recorded according to the International Society for Clinical Electrophysiology of Vision standards using a RETIport system (Roland Consult, Brandenburg, Germany) (21). The dark-adapted 3.0 ERGs were handled in a typical manner, with both a-wave and b-wave amplitudes analyzed.

Gene expression data (.tsv files) from postmortem eyes were downloaded from the GEO database (GSE135092), which compromising 128 macular retinal samples. Donors whose age were unknown were excluded for further analysis. A total of 57 normal control and nine AMD macular retinal samples from donors ranging from 75 to 90 years of age were included in the differential expression analysis. DEGs were identified with the DESeq2 package in R 4.1.2 and then ranked based on the simple p-values (<0.05). Gene ontology (GO) enrichment analysis and Gene set enrichment analysis (GSEA) were performed using the clusterProfiler and GGplot2 R package.

The paraffin-embedded retinal tissue sections were dewaxed to water, immersed in citric acid antigen repair buffer, boiled for antigen repair, and blocked with 10% goat serum (Beyotime, Jiangsu, China) for 2 h. Then the sections were incubated at 4 °C overnight with antibodies against glial fibrillary acidic protein (GFAP) (1:200; CST; #12,389), glutamine synthetase (GS) (1:250; Abcam; ab178422), and Aβ (1:200; Biolegend; #805,509).

Mice were sacrificed by cranio-cervical dislocation. Eyes were immediately enucleated and fixed in a sufficient amount of precooled 2.5% glutaraldehyde fixing solution (Servicebio, Wuhan, China) for 2 h. The anterior segments were removed, and eyecoat at the posterior pole was cut into square pieces measuring approximately 1 to 2 mm². After postfixing with 1% osmium tetroxide, dehydrating with graded acetone, embedding with resin, and slicing with an ultramicrotome (Leica Microsystems, Wetzlar, Germany), thin sections (50–100 nm) were examined on a HT7700 TEM (Hitachi, Tokyo, Japan).

Mice were sacrificed, and eyes were quickly enucleated. The cornea, iris, lens, and vitreous body were removed from the eye in cold PBS solution under a dissecting microscope (Olympus, Tokyo, Japan). Fresh retinas were lysed in RIPA buffer (NCM Biotech Co., Ltd., Suzhou, China) containing a protease inhibitor cocktail (APExBIO, Houston, TX, USA). The retinas were then ground with a cryogenic tissue grinder (Servicebio, Wuhan, China). After determination of the concentration of protein in each group using a BCA protein assay kit (Pierce; Thermo-Fisher Scientific, Waltham, MA, USA), the lysates were boiled. Proteins in the lysates were separated by 12% sodium dodecyl sulfate–polyacrylamide gel electrophoresis and subsequently transferred to PVDF membranes (EMD Millipore, Burlington, MA, USA). The membranes were blocked with 5% bovine serum albumin (NeoFroxx GmbH, Einhausen, Germany) for 1.5 h at room temperature and incubated overnight at 4 °C with primary antibodies. Antibodies against optic atrophy 1 (OPA1) (1:1 000; #80,471), dynamin-related protein 1 (DRP1) (1:1 000; #8,570), mitofusin 2 (MFN2) (1:1 000; #9,482), amyloid precursor protein (APP) (1:1 000; #76,600), and progesterone receptor membrane component 1 (PGRMC1) (1:1 000; #13,856) were obtained from Cell Signaling Technology (CST, Beverly, MA, USA). Antibodies against PTEN-induced kinase 1 (PINK1) (1:300; sc-517353) were obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Antibodies against transmembrane Protein 97 (TMEM97) (1:500; nbp1-30437) was purchased from Novus (Novus, St. Louis, MO, USA). An antibody against β-actin (1:2000; 66009-1-Ig) was purchased from Proteintech (Wuhan, China) and used as an internal control.

All statistical analyses were performed using GraphPad Prism 7.0 (GraphPad Software, CA, USA). For comparison between two groups, Student's t-test was used; for comparison between multiple groups, one-way ANOVA followed by Bonferroni's multiple comparison test were used. Data are presented as mean ± SEM. P-values < 0.05 were considered statistically significant.

Compared to drinking water, 100 mg/kg/day EC had no apparent toxic effects on the retina by day 7 (Figure 1). Fundus photography did not reveal any observable lesions in the EC group (Figure 1A). According to OCT (Figure 1B) or the HE-stained sections (Figure 1C), the oral EC treatment did not cause morphological changes in the retina. The average number of nuclei in the ONL was not significantly different between the two groups (Figure 1D). These results indicate that oral administration of EC is safe for retinal use in C57BL/6 mice.

To explore the retinal protective effects of EC, we established a mouse model of NaIO₃-induced retinal degeneration mimicking dry AMD by intravenous injection of 40 mg/kg NaIO₃. NaIO₃ caused disorganization of photoreceptors and sharp retinal thinning, especially in the ONL, by day 7. EC apparently ameliorated deformation of the outer retina (Figure 2A), partly decreased the total area of drusen-like deposits (although the trend was not statistically significant) (Figure 2B), increased the number of photoreceptor nuclei in the ONL (Figure 2C), and protected the ONL from severe NaIO₃-induced thinning (Figures 2D,E). However, differences between the NaIO₃ group and the EC group in the thickness of the IS/OS layer were not statistically significant (Figures 2F,G).

To further confirm the effectiveness of EC on a degenerative retina in vivo, we conducted non-invasive eye examinations. Color fundus photographs intuitively showed that exposure to NaIO₃ could induce diffuse dot-like foci or patchy yellow-white lesions in the retina, and oral gavage of EC significantly reduced those drusen-like deposits on the fundus (Figure 3A). The areas of degeneration were sharply reduced when used EC (Figures 3B,C). OCT images revealed that the outer retinal layers in eyes treated with systemic NaIO₃ were unclear, disorganized, and hyper reflective, but the hyper reflective foci were reduced when treated with EC (Figure 3D).

In addition to the effects on retinal morphology and structure, we then assessed how EC works on visual function. Dark-adapted 3.0 ERG showed that flat responses with both the a- and b-wave amplitudes were substantially decreased, by nearly 90% after the intravenous injection of NaIO₃ (Figure 4A). This indicates that combined responses arising from photoreceptors and second-order neurons were severely impaired. Although the a-wave amplitudes, which are physiologically generated from photoreceptors, did not show statistically significant changes with EC intervention (Figure 4B), an approximate 130 μV increase in b-wave amplitude was observed with the recovery effects of EC (Figure 4C).

To better explain molecular mechanisms underlying the protective effects of EC, we then identified key genes and pathways associated with AMD using gene expression profiling data of human retina samples (GSE135092). Compared to the normal controls, 548 DEGs in AMD samples were identified, with 344 genes upregulated and 204 genes downregulated (Figure 5A, Data S1). Among the top 100 genes with the most obvious expression changes (Data S2), TMEM97 (log₂FC = −0.637, p = 0.0079) and VTN (log₂FC = −0.502, p = 0.0491) are associated with AMD (22).

Enriched GO terms of interest are displayed in Figure 5B: DEGs were significantly enriched in biological processes such as response to oxidative stress (GO: 0006979), response to reactive oxygen species (GO: 0000302), astrocyte differentiation (GO: 0048708), glial cell migration (GO: 0008347), and Aβ clearance (GO: 0097242); and in molecular function such as Aβ binding (GO: 0001540) and antioxidant activity (GO: 0016209).

Additionally, DEGs that significantly enriched in pathways related to mitochondrial structure and function showed reduced expression in the AMD group compared to the control group in GSEA (Figure 5C).

Astrocyte and Müller cells regulate GFAP and GS expression in response to retinal injuries (23). Since genes involved in glial cell migration and astrocyte differentiation were enriched in GO analysis in AMD retinal samples, we studied retinal gliosis using GFAP and GS immunostaining in NaIO₃-induced retinal degenerative mice. Figure 6 shows that GFAP levels markedly increased with NaIO₃ administration, as expected. This enhancement was reduced by the EC administration (Figures 6A,B). The intensity of GS staining was sharply reduced in the NaIO₃-treated retinas as compared to the control retinas, but subsequent EC treatment showed an enhancement (Figures 6C,D).

Ultrastructural analysis verified the protective role of EC in preserving mitochondria in RPE cells. Results obtained by TEM showed abnormal mitochondrial morphology: reduced numbers of mitochondria, extensive disruption of the cristae, and decreased electron density of the matrix. Mitochondrial morphology and structure notably improved with the EC treatment (Figures 7A,B). Although EC did not change the mean area of mitochondria (Figure 7E), it increased the number of mitochondria per field (Figure 7C) and decreased the percentage of damaged mitochondria (Figure 7D) as compared to the NaIO₃ group.

To determine whether mitochondria-related pathways were involved in the protective effects of EC in NaIO₃-induced mitochondrial damage, we examined mitochondrial function in retina. First, the expressions of OPA1, MFN2, and DRP1, key proteins responsible for mitochondrial fission and fusion, were measured by Western blot (Figure 8A). The results showed that levels of OPA1, a protein responsible for fusion of inner mitochondrial membranes, were nearly normalized with EC administration (Figure 8B); levels of MFN2 and DRP1 did not change compared to NaIO₃ injected alone (Figures 8C,D). Interestingly, when treated with NaIO₃, levels of PINK1 markedly increased; when co-treated with EC, the levels of PINK1 were brought down (Figure 8E).

TMEM97, also known as sigma-2 receptor that important for the neuronal accumulation of Aβ (24, 25), is one of DEGs in AMD detected by bioinformatics analyses, so we explored the role of the TMEM97 and related proteins in this model using Western blot (Figure 9A). Interestingly, we found a notable increase in APP (Figure 9B) and TMEM97 protein levels (Figure 9C) and a decrease in PGRMC1 levels (Figure 9D) 7 days after NaIO₃ administration. EC treatment lowered this trend significantly. Because the molecular weight of Aβ is too small to be measured by Western blot, we detected its fluorescent intensity using immunofluorescence. The results indicated a trend toward a reduction of sub-RPE Aβ deposits in the EC group (Figures 9E,F).