The used animal was C57BL/6J male mice which were 7–8 weeks old in this study. All animals were maintained under standard laboratory conditions (room temperature 17–25°C, humidity 40–60%, 12 h dark/light cycle) with free food and water. All experiments with animals were approved by the Animal Research Ethics Committee of Zhengzhou University in accordance with administrative regulations for animal experimental in China. Project identification code was 2020-ky-67, and the date of approval was July 10, 2020. Much efforts were made to reduce the number of mice and their suffering. The initial density of MMS (M4882E961, ACMEC, China) is 1.3 g/ml. The injection method of MMS was single tail vein injection of 60 mg/kg, 5 μL/g body weight. The experiment of retinal degeneration induced by MMS was divided into two groups: Normal (n = 6) and MMS (1 day, n = 6; 2 day, n = 6; 3 day, n = 6; 4 day, n = 6; 5 day, n = 6; 7 day, n = 6; 10 day, n = 6). The injection method of rapamycin was intraperitoneal injection of 10 mg/kg, 10 μL/g body weight. Rapamycin was injected 1 day before MMS injection and every day until the fifth day after MMS injection. The experiment of rapamycin treatment was divided into four groups: Normal + PBS (n = 6), MMS + PBS (n = 6), Normal + Rapamycin (n = 6), and MMS + Rapamycin (n = 6).

All the mice were placed in the test room in advance, dark adapted for 12 h or more. Turn on the RETI-scan visual electrophysiological examination system (Roland Consult, Germany). The mice were placed in the isoflurane gas filled chamber to induce anesthesia. After the mice were anesthetized, they were fixed on the operating platform. The anesthesia mask was placed over their mouth and nose to maintain the deep anesthesia state. The breathing was stable, and the compound tropicamide eye drops were applied for mydriasis. Installation of electrodes: the grounding electrode was inserted into the subcutaneous tissue of mouse tail, two reference electrodes were inserted into the subcutaneous tissue of bilateral cheek, and two corneal ring electrodes were placed at the equator of eyeball, which could not compress eyeball. The mouse operating platform was pushed into the ganzfeld spherical stimulator to detect the five items of ERG proposed by the international society for clinical electrophysiology of vision (ISCEV).

The fixed eyeball was washed with pure water, placed in a plate, and operated under a stereo-microscope (Olypus/SZ61). The cornea, lens and iris tissues were removed, and the optic cup was retained. After conventional gradient alcohol dehydration, xylene transparency and wax immersion (SAKURA/VIP-5-J r-JC2), the eyeball was embedded in paraffin (SAKURA/TEC 5 EM JC-2). Paraffin tissue was cut into 4 μm thick sections (YAMATO/RX-860). Every three tissue sections were pasted on one slide, and three slides were prepared for each eyeball. The slices were baked at 65°C for 60 min, dewaxed with xylene, hydrated with gradient alcohol, soaked in pure water for 5 min, stained with hematoxylin for 20 s, differentiated with hydrochloric acid, blue returned with ammonia, stained with eosin for 10 s, dehydrated with gradient alcohol, transparent with xylene and sealed with neutral resin. Under the microscope (Olympus/BX53), the retina with 300 μm of myopic nerve was photographed at 400× and the parts of retina were photographed at 200× with Photoshop 5.0 software. The distance between the upper end of the outer retinal nuclear layer and RPE was measured by Image J software.

The mice were placed in the normal light of the test room for 2 h in advance. Open the computer’s topscan version 3.0 software for program settings. The open field area was divided into peripheral area, central area and total area, and the activity time of mice was set as 600 s per mouse. Each subject was placed in the center area of the open field apparatus. Total distance traveled (mm), time duration in the center area and peripheral area (s), velocity in the open field (mm/s) and bounts in the center area were recorded automatically. After testing a mouse, disinfect the test area with 75% alcohol and dry it, and then carry out the next target test.

The mice were placed in the normal light of the test room for 2 h in advance. Open the computer’s topscan version 3.0 software for program settings. The activity area was divided into Dark area and Light area, and the activity time of mice was set as 600 s per mouse. Each subject was placed in the Light chamber. Total distance traveled (mm), time duration in the Dark area and Light area (s), bounts in the Light area and proportion of time in black chamber and white chamber were recorded automatically. After testing a mouse, disinfect the test area with 75% alcohol and dry it, and then carry out the next target test.

Paraffin sections of retinal tissue were routinely dewaxed to water. Immunohistochemistry pen was used to circle around the retinal tissue. 20 μg/ml protease K without DNase (st532, Beyotime, China) was dripped. The cells were incubated at room temperature for 30 min and washed with PBS for 5 min × 3 times. Prepare the TUNEL detection solution (TDT enzyme: fluorescent labeling solution = 1:9), add 20 μL to each retina, put the slices into a wet box containing water, incubate them at 37°C for 60 min in dark, and wash them with PBS for 5 min × 3 times. Each tissue was incubated with 10 μl DAPI at room temperature for 3 min, washed with PBS for 5 min × 3 times, sealed with anti-fluorescence quenching solution, observed and took photos under fluorescence microscope.

The retina protein was extracted using RIPA buffer (P0013B, Beyotime, China) with PMSF and phosphatase inhibitors. The concentration of extracted protein was measured with BCA protein concentration assay kit (P0009, Beyotime, China). According to the measured value, the samples were adjusted to the same protein concentration (0.8 μg/μL) and denatured. 15 μL of each sample containing equal amounts of protein (12 μg) was taken for electrophoresis and transferred. The PVDF membrane with Western blotting was blocked with blocking solution (5% skimmed milk powder prepared with TBST) for 2 h, and then western blotting was incubated with primary antibody mTOR, 1:1,000 (Abcam), p-mTOR, 1:1,000 (Abcam), Rhodopsin, 1:500 (Abcam), and GAPDH, 1:10,000 (Prteintech) overnight at 4°C. After overnight, the membrane was washed three times with TBST and then incubated with goat anti-mouse IgG secondary antibodies 1:10,000 (Proteintech) and goat anti-rabbit IgG secondary antibodies 1:10,000 (Proteintech). After washed three times with TBST, the PVDF membrane was added with the ECL droplet to make it uniformly distribute on the surface of the whole membrane and was put into the gel imaging system to take photos.

Extracted the retina under the microscope, quickly cut the retina into 1 mm³ tissue and put it in the electron microscope fixing solution. Wash the tissue with 0.1M phosphate buffer PB (PH7.4) for three times, each time 15 min. The tissue was fixed with 1% OsO4 at room temperature and away from light for 2 h. The tissue was put in 30–50–70–80–95–100–100% ethanol for upward dehydration for 20 mins each time. Dehydrate with 100% acetone twice for 15 mins each time. Infiltrate the embedded tissue and overnight in a 37° oven. The embedded tissue was polymerized in a 60° oven for 48 h, and the resin block was used for slicing. Sliced the resin block and picked the piece with copper mesh. The sections were stained with citric acid and dried overnight. Took pictures with transmission electron microscope (HT7800/HT7700, Hitach) and saved the pictures.

Statistical analysis was performed using GraphPad Prism Version 7.0 software (Graph Pad software, San Diego, CA, United States). All measurement results are represented as Mean ± SD. One-way ANOVO was used for comparison of the effect of MMS at different time. Two-way ANOVO was used to evaluate the effect of rapamycin and MMS at the same time. P < 0.05 was considered statistically significant for all tests.

We took mouse eyeballs at day 1, day 2, day 3, day 4, day 5, day 7, and day 10 for fixation, paraffin embedding, paraffin sectioning and hematoxylin and eosin (HE) staining to obtain the morphological changes of retinal photoreceptor cells at different time points after MMS injection (Figure 1). The hemispheric section of retina is displayed (Figure 1A), and the position about 200 μm on both sides of the optic nerve is used to calculate the thickness of the retinal outer nucleus layer (ONL) where photoreceptor cells are located. The retinal cells were labeled in 40 fold HE sections (Figure 1B). One way ANOVO analysis (F = 616.9, P < 0.0001) revealed that the thickness of the ONL layer decreased at day 1, day 2, day 3, day 4, day 5, day 7, and day 10 after MMS injection compared with normal retina (Figure 1D), and the ONL was the thinnest 10 days after MMS injection, P < 0.0001 (Figures 1C,D). In addition to the gradual thinning of the ONL of the retina, RPE cells showed posterior vacuole structure at 4, 5, 6, and 7 days after MMS injection (asterisk), which also suggested that pigment epithelial cells also showed stress response with the damage of photoreceptor cells (Figure 1C).

We performed a full-field (Ganzfeld) flash electroretinogram (ERG) test at day 1, day 3, day 5, and day 7 to record the changes of retinal photoreceptor function (Figure 2). The amplitude of the ERG wave can show the function of the whole visual conduction circuit from retinal photoreceptor cells to retinal ganglion cells (Robson et al., 2018). The b-wave of dark adapted 0.01 ERG (flash strength is 0.01 photopic cd^▪s^▪m^–2 with a scotopic strength of 0.025 scotopic cd^▪s^▪m^–2) is mainly produced by rod bipolar cells in the retina, which is the main index to detect the function of rod cells. The b-wave of dark-adapted 3.0 ERG (flash strength is 3.0 photopic cd^▪s^▪m^–2 with a scotopic strength of 7.5 scotopic cd^▪s^▪m^–2) is mainly produced from the on-off path, which reflects the input of rod system and cone system. The a-wave of dark-adapted 3.0 ERG mainly produced from photoreceptor & post-receptoral on pathway, which reflects the function of photoreceptor (McCulloch et al., 2015). One way ANOVO analysis (F = 3,406; P < 0.0001) revealed that the amplitude of b-wave of dark-adapted 0.01 ERG decreased dramatically 1 day after Tail vein injection of MMS and the amplitude is the lowest at 7 days compared with normal retina, P < 0.0001 (Figure 2C). The amplitude of b-wave of dark-adapted 3.0 ERG decreased dramatically at 1 day after Tail vein injection of MMS and the amplitude is the lowest at 7 days compared with normal retina, P < 0.0001 (Figure 2D). Combined with the HE results of retina, we can conclude that the decline of retinal function was faster than the morphology changes of retina.

We speculated that MMS-induced retinal photoreceptor degeneration would seriously affect the visual acuity of mice and then their behavior. Therefore, we used open field test to detect the behavioral changes of mice at day 1, day 2, day 3, day 5, and day 7 after MMS injection (Figure 3). After MMS injection, the total distance of mice in the open field, the number of times (bounts) to the central area and the duration in the central area decreased compared with normal mice (Figures 3B–D), which showed that the changes of visual acuity in mice significantly affected their behavior. One way ANOVA multiple comparisons (F = 129.8, P < 0.0001) was to analyse the percentage duration of mice stay in the central area of the open field, P < 0.0001 for differences compared with controls; P < 0.001 for differences compared with previous time point group. One way ANOVA multiple comparisons (F = 80.77, P < 0.0001) was to analyse the duration of mice stay in the central area of the open field, P < 0.0001 for differences compared with Normal; P < 0.001 for differences compared with previous time point group. One way ANOVA multiple comparisons (F = 16.32, P < 0.0001) was to analyse the bounts of mice to the central area of the open field, P < 0.0001 for differences compared with Normal, P < 0.001 for differences compared with Normal. One way ANOVA multiple comparisons (F = 3.632, P < 0.05) was to analyse the total distance of mice in the open field, P < 0.05 for differences compared with Normal. It can be seen from the four indicators that the behavior of mice is the worst 3 days after MMS injection, and then there is a recovery, which suggests that the vision of mice may have a slight adaptation period.

TUNEL staining was used to measure the apoptotic process of retinal photoreceptors at day 1, day 2, day 3, day 5, and day 7 after MMS injection (Figure 4). We found that the apoptosis of retinal photoreceptor cells increased from the second day after MMS injection and reached the peak on the third day, while apoptotic cells had been basically eliminated on the seventh day (Figure 4).

Our results showed that the expression of rhodopsin which is the marker of retinal rod cells decreased gradually with the injection of MMS and was almost undetectable at 7 days (P < 0.0001), which displayed serious loss of rod cells (Figures 5A,B). We also detected that the expression of phosphorylated mTOR (p-mTOR) increased during retinal photoreceptor injury, which was obvious at 5 days (P < 0.001) and 7 days (P < 0.0001) after MMS injection (Figures 5A,C,D). This suggests that p-mTOR is activated during photoreceptor injury.

Because the expression of p-mTOR increased during MMS induced photoreceptor injury, we decided to use rapamycin, an inhibitor of mTOR to see whether it could reverse this injury. The experimental mice were divided into four groups with three mice in each group. The Normal + PBS group was only injected with PBS. The Normal + Rapa group was only injected with rapamycin. The MMS + PBS group was only injected with MMS. The MMS + Rapa group was injected with rapamycin intraperitoneally every day from the day before MMS injection until 5 days after MMS injection. Figure 6A showed the representative retinal morphology observed by H&E staining. Two-way ANOVA analysis revealed that both MMS (F = 185.2, P < 0.0001) and rapamycin (F = 11.78, P < 0.01) had effect on the thickness of the ONL, and there was a significant interaction between them (F = 18.72, P < 0.01). The thickness of ONL decreased in MMS + PBS group compared with the Normal + PBS group (P < 0.0001), while the thickness of ONL in MMS + Rapa group was significantly thicker than that in the MMS + PBS group (P < 0.01; Figures 6A,B), which exhibited the protective effect of rapamycin on the morphology of retinal photoreceptor. There was no significant difference in the thickness of ONL between the Normal + Rapa mice and the Normal + PBS mice (P > 0.05), indicating that rapamycin itself does not damage the retinal photoreceptor cells of mice.

There was no significant difference in the b-wave of Dark-adapted 0.01 ERG between the Normal + Rapa mice and the Normal + PBS mice (P > 0.05), indicating that rapamycin itself had no negative effect on retinal function of mice (Figures 6C,D). While the b-wave in Dark-adapted 0.01 ERG in the MMS + Rapa group was significantly higher than that in the MMS + PBS group (P < 0.05; Figures 6C,D), which displayed the protective effect of rapamycin on the function of mouse retinal rod cells. There was no significant difference in the amplitude of b-wave in Dark-adapted 3.0 ERG between the Normal + Rapa mice and the Normal + PBS mice (P > 0.05), indicating that rapamycin itself had no negative effect on retinal function of mice (Figures 6E,F). While the amplitude of b-wave in dark-adapted 3.0 ERG in the MMS + Rapa group was significantly higher than that in the MMS + PBS group (P < 0.01; Figures 6E,F), which displayed the protective effect of rapamycin on the function of mouse retinal photoreceptor. The absolute value of a-wave in dark-adapted 3.0 ERG between in the Normal + Rapa mice was bigger than that in the Normal + PBS mice (P < 0.01), indicating that rapamycin itself had positive effect on retinal photoreceptor function of mice (Figures 6E,G). Also, the absolute value of a-wave in dark-adapted 3.0 ERG in the MMS + Rapa group was significantly bigger than that in the MMS + PBS group (P < 0.01; Figures 6E,F), which displayed the protective effect of rapamycin on the function of mouse retinal photoreceptor. We also found that rapamycin could improve the retinal morphology and retinal function at 7 days after MMS treatment (Supplementary Figure 1).

The expression of rhodopsin in the MMS + Rapa group increased compared with the MMS + PBS group (P < 0.001), which was the molecular evidence for the improvement of retinal function and morphology (Figures 7A,B). The expression of p-mTOR in the MMS + Rapa group decreased compared to the MMS + PBS group (P < 0.05) while the expression of T-mTOR didn’t have significance difference (Figure 7D), which showed that the rapamycin inhibited the expression of p-mTOR (Figures 7A,C). We also detected autophagosome in the MMS + Rapa group by transmission electron microscope (TEM) (Figure 7E), which suggests that rapamycin may delay the death progress of retinal photoreceptor cells by inhibiting mTOR and then activating autophagy to remove the waste in the process of retinal photoreceptor cell death.

The experimental grouping was consistent with the above. There was no significant difference in the total distance and central duration between the Normal + Rapa mice and the Normal + PBS mice in the open field test (P > 0.05), indicating that rapamycin itself had no negative effect on the behavior the mice (Figure 8). While the total distance (P < 0.0001) and central duration (P < 0.01) of mice in the MMS + Rapa group was significantly longer than that in the MMS + PBS group (Figure 8), which displayed the positive effect of rapamycin on the behavior of the mice. There was no significant difference in the number of times to light area and the ratio of time in the dark area to time in the light area between the Normal + Rapa group and the Normal + PBS group in the open field test (P > 0.05), indicating that rapamycin itself had no negative effect on the behavior of the mice (Figure 9). The number of times to light area of mice in the MMS + Rapa group was significantly more than that in the MMS + PBS group (P < 0.05), while the ratio of time in the dark area to time in the light between these two groups showed no differences (Figure 9).