A total of 32 C57BL/6J pregnant mice (age, 14 weeks, weight 32-36 g) were purchased from Shanghai SLAC Laboratory Animal Co., Ltd. and maintained in our laboratory at a controlled temperature of 23±2°C and a relative humidity of 50±10% under a 12-h light/dark cycle with free access to food and water. For RGC isolation, approximately eight postnatal day 5 mice were used. Mice were anesthetized with a mixture of ketamine (80 mg/kg) and xylazine (4 mg/kg), and then euthanized in a chamber via 100% CO₂ exposure at a displacement rate of 50% volume/min for 50 min (17). After euthanasia, the eyes were immediately enucleated in cold dissection buffer containing HBSS (cat. no. 14025134; Gibco; Thermo Fisher Scientific, Inc.) and 2 mM HEPES (cat. no. H3784; MilliporeSigma). Subsequently, the retinas were separated from the sclera and pigment epithelium under a binocular dissecting microscope (MZ7.5; Leica Microsystems GmbH). The retinas were placed in dissection buffer and digested with 5 mg/ml papain for 0.5 h at 37°C. The suspensions of retinal cells were incubated with anti-macrophage antiserum (1:100; cat. no. AIA31240; Accurate Chemical & Scientific Corporation) for 1 h at 37°C to eliminate macrophages and microglial cells. The non-adherent cells were transferred to 100-mm petri dishes pre-conjugated with anti-Thy1.2 antibody (pre-conjugation for 3 h at room temperature; 1:20 dilution; cat. no. MCA02R; Bio-Rad Laboratories, Inc) to purify RGCs for 1 h at 37°C. The dishes were then rinsed with PBS to remove the non-adherent cells. The adherent cells were released using trypsin (cat. no. T4799; MilliporeSigma) and triturated with a pipette. RGCs were cultured in DMEM (cat. no. 21013024; Gibco; Thermo Fisher Scientific, Inc.) supplemented with insulin (2.0 mM; cat. no. I6634; MilliporeSigma), progesterone (40 nM; cat. no. P0130; MilliporeSigma), selenite (60 nM; cat. no. S5261; MilliporeSigma), transferrin (100 nM; cat. no. T8158; MilliporeSigma), CNTF (40 ng/ml; cat. no. 450-13; PeproTech, Inc.) and forskolin (6 µM; cat. no. F6886; MilliporeSigma) at 37°C and in an atmosphere containing 5% CO₂. For hypoxic induction, RGCs were cultured in a NAPCO 7001 incubator (The Precision Scientific Company) containing 1% O₂, 5% CO₂ and 94% N₂ at 37°C for 24 h. For oxidative stress induction, RGCs were exposed to H₂O₂ (50 µM) to mimic oxidative stress at 37°C for 24 h. The group without hypoxic induction or oxidative stress induction was taken as the control group.

The viability of RGCs was determined by CCK-8 assays (cat. no. CK04; Dojindo Laboratories, Inc.) according to the manufacturer's protocol. Briefly, 5×10³ RGCs/well were seeded into a 96-well plate. After transfection and exposure to hypoxic stress or oxidative stress, RGCs were incubated with 20 µl CCK-8 solution in each well for 3 h at 37°C. The absorbance was then detected at a wavelength of 450 nm using a microplate reader (Molecular Devices, LLC).

Caspase-3 activity was determined using the caspase-3 assay kit (cat. no. C1168S; Beyotime Institute of Biotechnology) according to the manufacturer's protocol. Briefly, RGCs were cultured at a density of 5×10³ cells/well in a 96-well plate. After the required treatment, cells were incubated with 100 µl caspase-3 substrate/well for 1 h at 37°C. Caspase-3 activity was determined under 490 nm excitation wavelength and 525 nm emission wavelength. Values were obtained from the blank group (caspase 3 substrate and cell culture medium, without RGCs) and subtracted from the experimental group values. The change in caspase-3 activity was calculated relative to the value obtained from the untreated cell group. Each sample was measured in duplicate.

A total of 35 male C57BL/6 mice (age, 8 weeks; weight, 25-30 g) were used for detecting circZYG11B expression change during retinal neurodegeneration and investigating the role of circZYG11B in retinal neurodegeneration. The mice were purchased from Shanghai SLAC Laboratory Animal Co., Ltd. The mice were housed in a pathogen-free environment (temperature, 23±2°C; humidity, 50±5%) under a 12-h light/dark cycle with free access to food and water. All operations were conducted under anesthesia via an intra-peritoneal injection of ketamine (80 mg/kg) and xylazine (4 mg/kg) (18,19). During I/R injury, the body temperature was maintained at 37°C using a temperature-controlled heating pad. The pupils were dilated using 0.5% phenylephrine and 0.5% tropicamide (Santen). Oxybuprocaine hydrochloride (Santen) was then applied to the corneas, followed by injection with a 30-gauge needle connected to a sterile bag containing 0.9% sodium chloride. To detect circZYG11B expression changes during I/R-induced retinal neurodegeneration, the anterior chamber of one eye was cannulated with a 30-gauge needle connected to the saline bag, which was located 120 cm above the eye, leading to a high IOP of ~90 mmHg. After removing the needle from the anterior chamber for 1 h, the IOP returned to normal. The contralateral eye was connected to the saline bag, which was not raised above the eye, thus a normal IOP was maintained as the control. The needle was withdrawn from the eye after 1 h. The retinas were collected at 0, 3 and 7 days after I/R injury (n=5 mice/time point).

To investigate whether silencing of circZYG11B affected retinal neurodegeneration, the mice were randomly separated into the following four groups: i) Control group: One eye was injected with the needle without elevating IOP and the other eye was left untreated (n=5 mice); ii) I/R group: One eye underwent I/R injury and the other eye was left untreated (n=5 mice); iii) I/R + control shRNA group: One eye was injected with scrambled (Scr) short hairpin (sh)RNA and underwent I/R injury and the other eye was left untreated (n=5 mice); and iv) I/R + circZYG11B shRNA: One eye was injected with circZYG11B shRNA and underwent I/R injury and the other eye was left untreated (n=5 mice). A total of 7 days after I/R injury, the mice were euthanized in a chamber via 100% CO₂ exposure at a displacement rate of 50% volume/min. The mice received an intravitreous injection of 1.0 µl (7.5×10⁻⁵ µg/ml) circZYG11B shRNA or equal amounts of Scr shRNA 14 days before I/R injury. The mice were injected once and sacrificed 21 days after injection. The sequence for circZYG11B shRNA was 5′-GGA CAC TTG AAG GAG GAA GCC TTC AAG AGA GGC TTC CTC CTT CAA GTG TCC TTT TTG-3′; the sequence for negative control shRNA was 5′-TCC TAA GGT TAA GTC GCC CTC GCT CGA GCG AGG GCG ACT TAA CCT TAG GTT TTT G-3′. The shRNAs (OBio Technology) were inserted into an AAV-2 vector (Stratagene; Agilent Technologies, Inc.) under the control of RNA polymerase III promoter U6, and then packaged with pHelper and pAAV-RC to produce recombinant AAV2-circZYG11B shRNA or AAV2-Scr shRNA. The target site for circZYG11B shRNA was located at the junctional site of circZYG11B spanning 9 nt sequence on one side and 12 nt on the other side.

RGCs were cultured onto 6-well plates at 37°C overnight at a density of 1×10⁵ cells/well. When cell confluence had reached 80%, RGCs were transfected with circZYG11B small interfering (si)RNA, negative control siRNA, Scr mimic, miR-620 mimic, pcDNA 3.1 (Vector) or PTEN-pcDNA 3.1 (all purchased from Sangon Biotech Co., Ltd.) using Lipo6000™ transfection reagent (cat. no. C0526; Beyotime Institute of Biotechnology) according to the manufacturer's protocol. circZYG11B siRNA, negative control siRNA, Scr mimic, miR-620 mimic, pcDNA 3.1 and PTEN-pcDNA 3.1 were diluted in Opti-MEM (Gibco; Thermo Fisher Scientific, Inc.), added to a culture plate and mixed gently 3-5 times. Lipo6000™ transfection reagent was also diluted in Opti-MEM and incubated for 5 min at room temperature. Subsequently, the diluted siRNA, mimics or plasmids, and transfection reagent were combined and incubated at room temperature for 20 min. The mixture was then added to each well for transfection. After 12 h at 37°C, the medium was replaced and cells were cultured at 37°C for an additional 12 h for the subsequent experiments. The sequences were as follows: circZYG11B siRNA, sense 5′-GGA CAC UUG AAG GAG GAA GCC TdT d-3′, anti-sense 5′-GG CUU CCU CCU UCA AGU GCU CCT dTd-3′; Scr siRNA, sense 5′-UUC UCC GAA CGU GUC ACG UTd Td-3′, anti-sense 5′-ACG UGA CAC GUU CGG AGA ATd TdT-3′; miR-620 mimic, 5′-AUG GAG AUA GAU AUA GAA AUU U-3′; negative control miRNA mimic, 5′-UCGCUUGGUGCAGGUCGGG-3′. During transfection, the concentration of miRNA mimics and siRNAs was 50 nM, and the concentration of plasmids was 4.0 µg.

Eyes obtained from the four groups of mice were enucleated and fixed in 4% paraformaldehyde at 4°C overnight. The eyes were then dehydrated in graded concentrations of ethanol and embedded in paraffin. Sections (5 µm) were cut vertically through the optic nerve head. Subsequently, the paraffin-embedded tissue sections were dewaxed with xylene and rehydrated in graded concentrations of ethanol. After washing with ddH₂O, the sections were stained with hematoxylin for 5 min at room temperature, stained with eosin for 45 sec at room temperature, and mounted with neutral resin at room temperature for long-term storage (cat. no. C0105S; Beyotime Institute of Biotechnology). The images were captured using an Olympus IX-73 bright-field microscope (Olympus Corporation).

The retinas were dissected in PBS and fixed in 4% paraformaldehyde for 12 h at room temperature. Subsequently, they were immersed in 30% sucrose solution for another 12 h and embedded in optimum cutting temperature compound mounting media. Sections (10 µm) were cut and placed on gelatin-coated slides. After rehydration with PBS, retinal sections were blocked with 5% bovine serum albumin (cat. no. ST025; Beyotime Institute of Biotechnology) for 30 min at 37°C to minimize nonspecific labeling. Retinal sections were incubated with primary antibodies against GFAP (1:200; cat. no. ab7260; Abcam), NeuN (1:300; cat. no. ab128886; Abcam) or anti-tubulin β3 (TUJ1; 1:300; cat. no. 801213; Biolegend, Inc.) overnight at 4°C. The slides were then washed with PBS containing 0.1% Tween 20 and incubated with Alexa Fluor 594 8 secondary antibodies (1:500; cat. nos. ab150080 and ab150116; Abcam) overnight at 4°C. After washing with PBS, the nuclei were labeled with DAPI (cat. no. D9542; MilliporeSigma) for 15 min at room temperature. The staining signaling was observed in six selected visual fields using an Olympus IX-73 fluorescence microscope (Olympus Corporation). Adobe Photoshop CS2 (Adobe Systems, Inc.) software was used for the preprocessing the fluorescent images.

The apoptosis of RGCs was determined using the TUNEL Apoptosis Detection kit (cat. no. C11026; Guangzhou RiboBio Co., Ltd.) according to the manufacturer's protocol. RGCs were fixed with 4% paraformaldehyde at 4°C for 30 min and were blocked with the blocking buffer (3% H₂O₂ in CH₃OH) at room temperature for 15 min. Subsequently, RGCs were permeabilized with 0.1% Triton X-100 for 10 min on ice, incubated with the TUNEL reaction mixture for 1 h at 37°C and stained with DAPI for 5 min at room temperature to label cell nuclei. The staining signaling was observed in six selected visual fields using an Olympus IX-73 fluorescence microscope (Olympus Corporation).

circZYG11B sequence was inserted into the KpnI and HindIII sites of pGL3 vector (Promega Corporation) to generate the Luc-circZYG11B vector. circZYG11B-interacting miRNAs were predicted using the circular RNA Interactome database (https://circinteractome.nia.nih.gov/) (20). RGCs were seeded in 96-well plates at a density of 5×10³ cells/well. After 24 h culture, RGCs were transfected with different miRNA mimics (final concentration: 50 nM) using Lipo6000™ transfection reagent according to the manufacturer's protocol. After 12 h, the medium was replaced and RGCs were cultured at 37°C for additional 12 h for luciferase activity assay. All miRNA mimics were synthesized by Sangon Biotech Co., Ltd. The mRNA mimics used were as follows: miR-149 mimic, 5′-UCU GGC UCC GUG UCU UCA CUC CCU U-3′; miR-186 mimic, 5′-CAA AGA AUU CUC CUU UUG GGC UUU-3′; miR-543 mimic, 5′-AAA CAU UCG CGG UGC ACU UCU UUU-3′; miR-545 mimic, 5′-UCAGCA AAC AUU UAU UGU GUG CUU-3′; miR-568 mimic, 5′-AUG UAU AAA UGU AUA CAC ACU U-3′; miR-607 mimic, 5′-GUU CAA AUC CAG AUC UAU AAC UU-3′; miR-616 mimic, 5′-AGU CAU UGG AGG GUU UGA GCA GUU-3′; miR-587 mimic, 5′-UUU CCA UAG GUG AUG AGU CAC UU-3′; miR-620 mimic, 5′-AUG GAG AUA GAU AUA GAA AUU U-3′; miR-624 mimic, 5′-CAC AAG GUA UUG GUA UUA CCU UU-3′; miR-634 mimic, 5′-AAC CAG CAC CCC AAC UUU GGA CUU-3′; miR-654 mimic, 5′-UGG UGG GCC GCA GAA CAU GUG CUU-3′; miR-936 mimic, 5′-ACA GUA GAG GGA GGA AUC GCA GUU-3′; miR-95 mimic, 5′-UCA AUA AAU GUC UGU UGA AUU UU-3′; The luciferase activity was detected using a Dual-Luciferase Reporter Assay kit (Promega Corporation). The vector pRL-TK expressing Renilla luciferase was used as the internal control for transfection. The empty vector pGL3-basic was used as the negative control. For comparisons, the luciferase activity of the pGL3 vector was normalized with Renilla luciferase activity.

RGCs (~1×10⁶) were transfected with biotinylated miR-620 mimics (5′-AUG GAG AUA GAU AUA GAA AUU U-3′) or control mimics (miR-95 mimic: 5′-UCA AUA AAU GUC UGU UGA AUU UU-3′), which were synthesized by Sangon Biotech Co., Ltd., using Lipo6000™ transfection reagent according to the manufacturer's protocol. The final concentration of miRNA mimics was 50 nM. After transfection for 24 h, the cells were lysed in 500 µl lysis buffer (50 mM Tris-HCl, pH 7.5; 150 mM NaCl; 1% Triton X-100; 1% sodium deoxycholate; 0.1% SDS). Subsequently, 50 µl washed streptavidin magnetic beads (Thermo Fisher Scientific, Inc.) were blocked in 1% bovine serum albumin (cat. no. ST025; Beyotime Institute of Biotechnology) for 2 h (21). These beads were then added to the reaction tubes and incubated with 300 µl lysates for 3 h at 4°C to pull down the biotin-coupled RNA complex. After washing the beads with ice-cold lysis buffer, RNA was isolated with TRIzol^® reagent (cat. no. 15596018; Invitrogen; Thermo Fisher Scientific, Inc.) and the expression levels of circZYG11B and circZNF236 (negative control) were detected by reverse transcription-quantitative PCR (RT-qPCR).

RIP assays were conducted using the EZ-Magna RIP kit (MilliporeSigma) according to the manufacturer's protocol. Briefly, RGCs were lysed in the RIP lysis buffer and the cell lysate was incubated with magnetic beads conjugated with anti-Ago2 or anti-IgG antibody for 8 h at 4°C. The beads were then incubated with proteinase K (Invitrogen; Thermo Fisher Scientific, Inc.) for 30 min at 58°C to remove proteins. Finally, the immunoprecipitated RNAs were extracted using TRIzol reagent according to the manufacturer's instructions and subjected for RT-qPCR to determine the expression levels of circZYG11B.

The target gene information of miR-620 was predicted using TargetScan software (http://www.targetscan.org/) (22).

Total RNAs were extracted from retinal tissues or RGCs using TRIzol reagent according to the manufacturer's instructions. For mRNA expression detection, total RNAs were reverse transcribed using the SuperScript III first-strand synthesis kit (Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. The temperature protocols for RT were as follows: 37°C for 60 min, 95°C for 5 min and maintenance at 4°C. Subsequently, qPCR assays were conducted using the PowerUp™ SYBR™ Green Master Mix (cat. no. A25742; Applied Biosystems; Thermo Fisher Scientific, Inc.) in a PikoReal 96 cycler (Thermo Fisher Scientific, Inc.) using gene-specific primers (Table I) according to the manufacturer's protocol. The thermocycling protocols were as follows: Initial denaturation at 95°C for 3 min, followed by 42 cycles at 95°C for 10 sec, 56°C for 30 sec and 72°C for 30 sec, followed by final extension at 72°C for 60 sec. GAPDH was used as a housekeeping gene.

For miR-620 detection, RT was conducted using the miScript Reverse Transcription kit (Qiagen GmbH) according to the manufacturer's protocol in a final volume of 20 µl containing 5 µl total RNA, 5 µl 5X miScript RT buffer and 1 µl miScript Reverse Transcriptase Mix. The temperature protocols for RT were as follows: 37°C for 60 min, at 95°C for 5 min and maintenance at 4°C. qPCR was conducted using the miScript SYBR-Green PCR kit (Qiagen GmbH) in a PikoReal 96 cycler (Thermo Fisher Scientific, Inc.) using gene-specific primers (Table I) according to the manufacturer's protocol. The thermocycling protocols were as follows: Initial denaturation at 95°C for 3 min, followed by 42 cycles at 95°C for 10 sec, 56°C for 30 sec and 72°C for 30 sec, followed by final extension at 72°C for 60 sec. Small nuclear RNA U6 was used as a housekeeping gene. Relative mRNA or miR-620 expression was calculated using the 2^−ΔΔCq method (23).

Aqueous humor (AH) samples were obtained from Eye & ENT Hospital (Shanghai, China) between November 2018 and March 2019. Written informed consent was obtained from the patients involved. Glaucoma AH samples were obtained from patients diagnosed with acute angle-closure glaucoma (12 male and eight female patients; age range, 53-66 years). Control AH samples were obtained from patients diagnosed with cataracts without acute angle-closure glaucoma (11 male and nine female patients; age range, 50-63 years). The patients were diagnosed at the hospital through ocular examination, including anterior segment photography, Goldmann applanation tonometry, fundus examination, visual field, optic disc photography and ultrasound biomicroscopy examination. The patients with glaucoma were diagnosed as having acute angle-closure glaucoma without cataracts and required filtration surgery. Patients were excluded if they had other ophthalmic or systemic diseases, had accepted systemic and local steroids therapy for glaucoma treatment, or had accepted ophthalmic surgery or intravitreous injections of anti-VEGF drugs.

The levels of PTEN in AH samples were determined using a human PTEN ELISA Kit (cat. no. ab206979; Abcam) according to the manufacturer's protocol. Briefly, the standard samples and AH samples were added to the pre-coated 96-well plates and incubated for 1 h at room temperature, followed by the addition of the antibody cocktail to each well. Subsequently, the plates were sealed and incubated for 1 h at room temperature. After washing the plates three times, the TMB Development Solution was added to each well and incubated for 10 min in the dark on a plate shaker set to 400 rpm. Finally, the Stop Solution was added to each well to terminate the enzyme-substrate reaction. After the reaction was terminated for 3 min, the absorbance value of each well was measured at a wavelength of 450 nm. A standard curve was plotted for determining the concentration of PTEN in each sample. All measurements were performed in duplicate.

All data were analyzed using GraphPad Prism 5 (GraphPad Software, Inc.). The experiments were independently repeated three times and the data are presented as the mean ± standard deviation. Comparisons between two independent samples were analyzed using the paired Student's t-test (Comparison of circZYG11B expression between I/R retinas and normal retinas) or unpaired Student's t-test (comparison of circZYG11B, miR-620 and PTEN levels in AH samples between patients with glaucoma and patients with cataracts), and multiple comparisons were made using one-way ANOVA followed by Bonferroni's test. P<0.05 was considered to indicate a statistically significant difference.

Models of retinal I/R injury are important for studying retinal neurodegeneration (16). Retinal I/R injury was induced by an acute increase in IOP. The present study used a mouse model of retinal I/R injury and detected the expression levels of circ-ZYG11B between I/R retinas and control retinas. IOP in one eye was raised to ~90 mm Hg for 1 h, and the contralateral eye was cannulated and maintained at normal IOP as the control. The needle was withdrawn from the eye after 1 h. The mice were sacrificed at 0, 3 and 7 days after I/R injury. The results demonstrated that circZYG11B was significantly increased in the retinas at day 3 or day 7 after I/R injury (Fig. 1A). By contrast, the mRNA expression levels of ZYG11B were not altered in the retinas after I/R injury (Fig. 1B). Retinal neurons, particularly RGCs, have been shown to be highly sensitive to I/R injury (3). In the present study, the isolated RGCs were exposed to H₂O₂ (50 µM) or 1% O₂ at 37°C for 24 h to mimic oxidative stress or hypoxic stress. The group without exposure to H₂O₂ or 1% O₂ was taken as the control group. The results revealed that compared with in the control group, exposure to H₂O₂ or 1% O₂ increased the expression levels of circZYG11B. By contrast, exposure to H₂O₂ or 1% O₂ did not alter the mRNA expression levels of ZYG11B (Fig. 1C and D).

The present study then investigated the role of circZYG11B in RGCs in vitro. RT-qPCR confirmed that transfection with circZYG11B siRNA decreased the expression levels of circZYG11B compared with in the control group (Fig. 2A). Hypoxic stress is an important pathological factor in retinal I/R injury (24). The non-transfected RGCs, or RGCs transfected with circZYG11B siRNA or Scr siRNA were exposed to 1% O₂ to mimic hypoxic stress in vitro. The non-transfected RGCs without 1% O₂ treatment was taken as the control group. CCK-8 assays revealed that compared with in the control group without 1% O₂ exposure, hypoxic stress induced by 1% O₂ led to decreased RGC viability. Compared with the 1% O₂ group, transfection of circZYG11B siRNA, but not Scr siRNA, significantly alleviated hypoxic stress-induced reduction of RGC viability (Fig. 2B). TUNEL assays revealed that compared with the control group without 1% O₂ exposure, 1% O₂ treatment led to an increased number of apoptotic RGCs. The number of apoptotic RGCs in the circZYG11B group was lower than that in the 1% O₂ group or in the 1% O₂ + Scr siRNA group (Fig. 2C and D). Caspase-3 activity assays revealed that hypoxic stress enhanced the activity of caspase-3. Compared with in the 1% O₂ group, transfection with circZYG11B siRNA, but not Scr siRNA, reduced the activity of caspase-3 induced by hypoxic stress (Fig. 2E). Collectively, silencing of circZYG11B may have a protective role in RGCs against hypoxic stress.

Oxidative stress is another pathological factor in retinal I/R injury (24). The non-transfected RGCs or RGCs transfected with circZYG11B siRNA or Scr siRNA were exposed to H₂O₂ (50 µmol/l) to mimic oxidative stress in vitro. CCK-8 assays revealed that compared with in the control group without H₂O₂ exposure, H₂O₂ treatment-mediated oxidative stress led to reduced viability of RGCs. Compared with in the H₂O₂ group, transfection with circZYG11B siRNA, but not Scr siRNA, alleviated the oxidative stress-induced reduction of RGC viability (Fig. 3A). TUNEL assays revealed that compared with in the control group without H₂O₂ exposure, treatment with H₂O₂ (50 µmol/l) led to an increased number of apoptotic RGCs. However, the number of apoptotic RGCs was significantly decreased post-transfection with circZYG11B siRNA, but not Scr siRNA (Fig. 3B). Oxidative stress led to enhanced caspase-3 activity in RGCs. Compared with in the H₂O₂ group, transfection with circZYG11B siRNA, but not Scr siRNA, significantly decreased the activity of caspase-3 induced by H₂O₂ (Fig. 3C). These findings suggested that silencing of circZYG11B could exert protective effects on RGCs against oxidative stress.

The present study next determined the role of circ-ZYG11B in I/R-induced retinal injury in vivo. C57BL/6 mice received an intravitreous injection of circZYG11B shRNA to knockdown the expression of circZYG11B. Injection with circZYG11B shRNA significantly reduced the expression levels of circZYG11B compared with in the control group, but did not alter the mRNA expression levels of ZYG11B in the retinas (Fig. 4A and B). H&E staining revealed that silencing of circZYG11B could partially reverse the reduction in I/R injury-induced retinal thickness and reduce I/R-induced neurodegeneration in the ganglion cell layer (GCL) (Fig. 4C and 4D). Retinas were also extracted and the surviving RGCs were quantified at day 7 after I/R injury. The density of surviving RGCs in the I/R group was decreased by ~40% compared with in the control group. Silencing of circZYG11B exerted protective effects on RGC survival and protected against I/R-induced RGC injury (Fig. 4E and F).

The present study further determined the role of circZYG11B in I/R-induced retinal degeneration in vivo. Immunofluorescence staining with NeuN was conducted to identify RGCs in the GCL. Compared with in the control retinas, I/R injury led to reduced expression of NeuN. Compared with in the Scr shRNA-injected retinas, injection of circZYG11B shRNA alleviated I/R-induced RGC injury as shown by increased NeuN staining (Fig. 5A and B). The retinas were then stained with GFAP to detect retinal reactive gliosis. Compared with in the control retinas, I/R injury led to activation of retinal reactive gliosis. Silencing of circZYG11B could alleviate I/R-induced retinal reactive gliosis as shown by decreased GFAP staining (Fig. 5C and D). These findings indicated that circZYG11B silencing could alleviate I/R-induced retinal neurodegeneration.

The present study next investigated the underlying mechanism of circZYG11B in retinal neurodegeneration. The Ago2 protein is a core component of the RNA-induced silencing complex that binds miRNA complexes to target genes. miRNAs associate with the Ago2 protein to post-transcriptionally suppress target gene expression (20). The RIP assays indicated that circZYG11B could be pulled down by Ago2 antibody but not IgG, suggesting that circZYG11B may serve a role by acting as a miRNA sponge (Fig. 6A). Luciferase activity assay demonstrated that transfection with miR-620 mimics or miR-654 mimics, but not the other miRNA mimics, reduced the activity of Luc-circZYG11B (Fig. 6B). Due to the greater silencing efficiency of miR-620 compared with miR-654 on the activity of Luc-circZYG11B, the present study focused on the role of miR-620 in RGCs. Because miR-95 mimics did not reduce the luciferase activity of Luc-circZYG11B, it was suggested that miR-95 could not directly bind to circZYG11B; therefore, miR-95 was selected as the negative control for RNA pull-down assays. RNA pull-down assays revealed that circZYG11B but not circZNF236 was greatly enriched in the miR-620-captured fraction. By contrast, there was no difference between the amount of circZYG11B and circZNF236 in the miR-95-captured fraction (negative control) (Fig. 6C). Subsequently, the potential target genes of miR-620 were predicted using TargetScan software (http://www.targetscan.org/) (22). The candidate gene, PTEN, was assessed further due to its role in cell apoptosis.

RT-qPCR revealed that transfection with a miR-620 mimic led to increased expression of miR-620 but decreased expression of PTEN (Figs. S1 and 6D). The present study then investigated the role of miR-620 in RGC biology in vitro. All experimental groups were exposed to 1% O₂ to mimic hypoxic stress. The non-transfected group was taken as the control group (hypoxia group). CCK-8 and TUNEL assays demonstrated that compared with in the control group, transfection with the miR-620 mimic, but not Scr mimic, led to increased cell viability and reduced number of TUNEL-positive cells, thus suggesting that the miR-620 mimic exerted protective effects on RGC function (Fig. 6E and F). Western blotting demonstrated that transfection with pcDNA3.1-PTEN increased the protein expression levels of PTEN in RGCs (Fig. S2). Notably, overexpression of PTEN could interrupt the protective effects of miR-620 on RGC function (Fig. 6E and F). Collectively, these results suggested that circZYG11B may regulate RGC function via the circZYG11B/miR-620/PTEN signaling axis.

The present study also investigated the clinical significance of circZYG11B-mediated signaling. Acute angle-closure glaucoma is an I/R-related ocular disease (23). AH samples were collected from patients with glaucoma (n=20) and patients with cataracts (control; n=20). circZYG11B expression was significantly higher in the AH samples from patients with glaucoma compared with that in control patients (Fig. 7A). By contrast, miR-620 expression was significantly downregulated in the AH samples from patients with glaucoma compared with control patients (Fig. 7B). The concentrations of PTEN in AH samples were determined using ELISA. The results revealed that PTEN concentration was markedly increased in the AH samples of patients with glaucoma compared with the control patients, showing a similar pattern as circZYG11B (Fig. 7C). Collectively, these results suggested that circZYG11B-mediated signaling is potentially involved in I/R-related ocular disease.