Experimental animals

We purchased C57BL/6 J male mice (6–8 weeks old) from Beijing SPF Biotechnology Co., Ltd. (Beijing, China). The Guidelines for the Care and Use of Laboratory Animals and the Principles for the Utilization and Care of Vertebrate Animals were strictly followed. The Institutional Animal Care and Use Committee of Fujian Medical University approved all animal experiments (IACUC FJMU 2021-0454). The procedures conformed to the guidelines of the Association for Research and Vision and Ophthalmology statement for the use of animals in ophthalmic and vision research. Streptozotocin (STZ; Sigma-Aldrich, USA) mixed with citrate-citric acid buffer was prepared for induction of type 1 diabetes mellitus (T1DM). Experimental mice (n = 65) were injected intraperitoneally with low-dose STZ (50 mg/kg) for five days, and control mice (n = 45) were injected intraperitoneally with an equal amount of citrate-citric acid buffer. Before intraperitoneal injection, we measured baseline blood glucose levels in both groups of mice, and then random intravenous blood glucose concentrations were measured every four weeks until 16 weeks. Both groups of mice were given a normal diet. Experimental mice whose blood glucose level exceeded 16.7 mmol/L in each measurement were considered to be diabetic.

Corneal epithelial debridement healing

An intraperitoneal injection of 1.25% tribromoethanol (0.2 mL/10 g) was used to induce general anesthesia in experimental mice, followed by local anesthesia of the ocular surface with 0.5% promethazine hydrochloride. An AlgerBrush II ring drill (Alger Inc., USA) was applied to exfoliate the corneal epithelium (2.5 mm), and then ofloxacin eye ointment was applied to prevent infection (Fig. 1a). At 0, 12, 24 and 36 h after debridement, experimental mice were grasped to briefly expose the eyeballs after local anesthesia, the defect area was observed with 0.25% sodium fluorescein staining and photographed under a microscope with cobalt-blue light. Image J was used to calculate the ratio of fluorescent stained area to corneal area as the percentage of epithelial defect area. At least three mice in each group were used for separate independent experiments.

Fig. 1

Sustained hyperglycemia delayed corneal debridement healing in mice. a Flowchart for the establishment of type 1 diabetes mellitus (T1DM) mice and corneal debridement healing. b Blood glucose values of normal (Ctrl) and diabetic (DM) mice. b1 Blood glucose values at 0, 4, 8, 12, 16 weeks after intraperitoneal injection. b2 Significant differences in blood glucose levels between the two groups at 16 weeks. c Body weight values of Ctrl and DM mice at 16 weeks. d Corneal sensitivity of Ctrl and DM mice at 16 weeks. e Corneal nerves whole-mount staining on day 5 of Ctrl and DM mice after debridement (n = 6 per group). e1 Fluorescent images of corneal nerve staining. e2 Peripheral corneal nerve density. f Corneas stained with fluorescein sodium at 0, 12, 24, and 36 h in Ctrl and DM mice after debridement (n = 6 per group). f1 Fluorescein-stained images of corneas. f2 Percentage of epithelial defect area. *, P < 0.05; ***, P < 0.001; ****, P < 0.0001

Corneal sensitivity measurement

The Cochet-Bonnet esthesiometer (Luneau ophthalmology, France) was used to estimate corneal sensitivity in this study. On day 5 after corneal epithelial debridement, each eye of unanesthetized mice was measured at least three times, and the blink reflex was judged as a positive response. Each measurement was shortened by 0.5 cm until a positive response was observed. The corneal sensitivity threshold is the longest filament length that results in a positive response. At least three mice in each group were used for separate independent experiments.

Whole-mount staining of corneal nerves

Experimental mice were sacrificed and fresh intact corneal tissue was harvested and fixed in Zamboni fixative on day 5 after corneal epithelial debridement. Corneal tissue was subsequently blocked and permeabilized in a solution containing 0.3% Triton X-100 and 10% goat serum. Anti-β-tubulin III antibody (657404, Biolegend, USA) was used for overnight staining. Corneal tissue was scanned under a laser confocal microscope (Leica, Germany). At least three mice in each group were used for separate independent experiments.

Western blot assay

The TG tissues were harvested on day 5 after corneal epithelial debridement and lysed with RIPA reagent (Beyotime, Shanghai, China) supplemented with 1% Phenylmethylsulfonyl Fluoride (PMSF) reagent (Beyotime). Protein samples were electrophoresed on Sodium dodecyl sulfate–polyacrylamide gels, transferred to polyvinylidene fluoride membranes, and blocked with skim milk. Afterward, the membranes were incubated in solutions of the corresponding primary antibodies and horseradish peroxidase-conjugated secondary antibodies. Eventually, Enhanced chemiluminescence was used to visualize proteins bands. At least three independent experiments were performed. The primary antibodies used for the study were as follows: anti-p62 antibody (p0067; Sigma-Aldrich, USA), anti-LC3B (ab192890; Abcam, USA), anti-ATG4D (ab237751; Abcam, USA), and anti-β-actin (ab8226; Abcam, USA).

Subconjunctival injection

After general and ocular surface anesthesia of mice, 5 µL of solution was injected into the lower bulbar conjunctiva with a microsyringe. For the injection of two solutions, different solutions were injected into the upper and lower bulbar conjunctiva. At least three mice in each group were used for separate independent experiments. All solutions were injected 24 h before, 0 h, and 24 h after debridement. The solutions for injection were as follows: RAPA solution (10 μmol/L) (HY-10219, MedChemExpress), 3-MA solution (20 μmol/L) (HY-19312, MedChemExpress). miR-542-3p agomir and miRNA agomir negative control (200 nmol/L; Ribobio, Guangzhou, China), miR-542-3p antagomir and miRNA antagomir negative control (200 nmol/L; Ribobio, Guangzhou, China), ATG4D antisense oligonucleotide (ATG4D ASO, 200 nmol/L; Ribobio, Guangzhou, China).

Immunofluorescence

The TG tissues were removed from experimental mice on day 5 after corneal epithelial debridement, embedded in optimal cutting temperature compound, and stored at − 80 °C. Frozen sections were cut at a thickness of 5 μm. After fixing the sections with 4% paraformaldehyde, the sections were permeabilized and blocked, then incubated with the corresponding primary antibodies and Alexa Fluor-labeled secondary antibodies. Next, the tissues were counterstained with DAPI to visualize the cell nuclei. Finally, the fluorescence intensities were estimated using a fluorescence microscope (Leica, Germany). At least three independent experiments were performed. The primary antibodies used were as follows: anti-p62 antibody (P0067; Sigma-Aldrich, USA), anti-LC3B (ab192890; Abcam, USA), and anti-ATG4D (ab237751; Abcam). The imaging parameters of the same protein were consistent for all experiments (Additional file 1: Table S1). 

Transmission electron microscopy

TG tissue were removed and cut into 1 mm3 sections then immersed in electron microscope fixative and 1% osmic acid. Using gradient ethanol and acetone solutions, the fixed tissues were dehydrated for 15 min. Subsequently, the tissues were embedded in epoxy resins and baked in an oven at 60 °C for 48 h, then cut into ultrathin Sections (80 nm) with an ultramicrotome. Sections were dried after staining with 2% double uranium-lead citrate. Transmission electron microscopic (HT-7700; Hitachi LTD., Japan) images were acquired to determine the number and volume of autophagosomes.

miRNA sequencing and data analysis

Three mice from each of the control and diabetic groups were collected, and the TG tissue were subjected to miRNA sequencing. Libraries were constructed, and the quality of the amplified libraries was verified with an Agilent 2100 Bioanalyzer (USA). Fifty sequencing cycles were performed using an Illumina NextSeq 500 (Illumina, USA). miRNAs with mean counts per million reads above one in each group were included in the statistical analysis, and paired samples were screened for differentially expressed miRNAs (DEmiRNAs) using the edgeR analytic tool. Only DEmiRNAs with P < 0.05, |log2 fold change|≥ 0.585 were included in the subsequent studies (Additional file 1: Tables S2 and S3).

TargetScan and miRDB are databases for miRNA target prediction. We predicted the target genes of miRNAs, from which we screened the target genes with a target score of at least 70. Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis of target genes was conducted to screen unwanted autophagy pathway-related mRNAs, and Cytoscape was used to visualize the animal autophagy pathway-related miRNA-mRNA regulatory network.

Quantitative real-time polymerase chain reaction (qRT-PCR)

The PrimeScript™ First Strand cDNA Synthesis kit (Takara, Japan) and miRNA First Strand cDNA Synthesis kit (Tailing Reaction) (Sangon Biotech, Shanghai, China) were utilized for cDNA synthesis. cDNA was amplified with qRT‒PCR with BlasTaq™ 2X qPCR MasterMix (Applied Biological Materials Inc., Canada) on an ABI7500 Real-Time PCR system (Applied Biosystems, Singapore), and the expression levels were analyzed with the 2−ΔΔCt comparison method using β-actin and U6 as internal references. Sangon Biotech (Shanghai) Co., Ltd. designed the primers in this study (Additional file 1: Table S4). At least three independent experiments were performed.

Dual-luciferase gene reporter assay

To clarity the interaction between mmu-miR-542-3p and ATG4D, plasmid vectors containing wild-type or mutant ATG4D were first constructed (Hanbio, Shanghai, China). Thereafter, the wild-type or mutant ATG4D plasmid was cocultured with cells transfected with miR-542-3p or negative control (NC), and luciferase activity was detected. Three independent experiments were performed.

Evaluation of ocular surface toxicity

miR-542-3p antagomir is the chemically-modified reverse complementary strand of miR-542-3p: 2 phosphorothioates at the 5’ end, 4 phosphorothioates at the 3’ end, 3’ end cholesterol group, and full-length nucleotide 2’-methoxy modification. We configured miR-542-3p antagomir powders (Ribobio, Guangzhou, China) with sterilized ddH2O to 200 nmol/L for subconjunctival injection. Experimental mice were injected with 5 µL solution subconjunctivally every two days for one week. During this period, intraocular pressure (IOP) was measured with a Tonolab tonometer (icare, Shanghai, China) on days 0, 1, 3, 5, and 7 (averaged three times). Corneal transparency and corneal epithelial integrity were evaluated under a slit lamp on day 7. Afterward, corneas were embedded in paraffin and sectioned (thickness 3 µm) for hematoxylin and eosin staining to measure corneal thickness with an optical microscope (Leica, Germany). At least three mice in each group for were used separate independent experiments.

Statistical analysis

Experiments were conducted at least three times independently. GraphPad Prism v.9.5.1 software (GraphPad Software, Inc., San Diego CA) was used to statistically analyze the data. Data were expressed as the mean ± standard error of the mean (SEM) and were compared between groups using an unpaired t test or one-way analysis of variance (ANOVA). Pearson coefficients were calculated to evaluate the correlation between the two groups of variables. A P value < 0.05 was considered statistically significant.