Tg(Lyz:Dsred) zebrafish were purchased from the China Zebrafish Resource Center (CZRC, ID: CZ59, China). Tg(isl2b.2:Gal4-VP16,myl7:EGFP);Tg(4XnrUAS:GFP) were kindly provided by Professor Jiulin Du (Institute of Neuroscience, State Key Laboratory of Neuroscience, Center for Excellence in Brain Science and Intelligence Technology, Chinese Academy of Sciences, Shanghai, China). Zebrafish were maintained under standard conditions (14 h light/10 h dark cycle, 28.5 °C), with embryos [0–7 days after fertilization (dpf)] reared in E3 medium (5 mM NaCl, 0.17 mM KCl, 0.33 mM CaCl2, 0.33 mM MgSO4). They were fed in the morning and evening and housed at the Eye Hospital of Wenzhou Medical University. To suppress melanocyte pigmentation, zebrafish larvae were maintained in E3 medium containing 0.2 mM PTU (Sigma-Aldrich, USA) at 28.5 °C starting at 1 dpf.

The larvae (5 dpf) were fixed in Bouin’s solution (Phygene, China) at 4 °C for 12 h then rinsed and dehydrated through a series of ethanol, xylene, and paraffin solutions using a HistoCore PEARL processor (Leica, Germany). Following dehydration, the larvae were embedded in paraffin with their heads facing downward and sectioned at 3 μm thickness using a HistoCore BIOCUT (Leica, Germany).

The sections were dried at 65 °C for 1 h. Dewaxing and H&E staining were performed using an Autostainer XL (Leica, Germany). Tissue sections were deparaffinized in xylene for 8 min, rehydrated through graded ethanol (100% for 6 min, 70% for 2 min), and rinsed in water for 2 min. Hematoxylin staining was performed for 3 min, followed by differentiation (45 s), bluing (1 min), and eosin counterstaining (1 min). Slides were dehydrated in ethanol (95% for 2 min, 100% for 4 min) and cleared in xylene for 4 min. The images were acquired and analyzed using the 3D Pannoramic Scanner Software (3DHISTECH, Hungary).

Larvae were embedded with the dorsal side up in 1.5% low-melting-point agarose (Sigma-Aldrich, USA) prepared in E3 medium. Using a SZ61TR microscope (Olympus, Japan), the right eye was excised at the proximal exit site with sharpened insect pins. After excision, larvae were released from the agarose and allowed to recover for 1 h in Ringer’s solution (38.7 mM NaCl, 1.0 mM KCl, 1.7 mM HEPES, 2.4 mM CaCl2, pH 7.2) supplemented with 0.2 mM PTU. Following recovery, larvae were transferred to E3 medium containing 0.2 mM PTU. After 24 h, larval viability was confirmed, and the optic nerve was transected for subsequent imaging or OKR experiments.

Optic nerve transection was performed on zebrafish larvae at 5 dpf, as previously described [8]. In brief, larvae were immobilized ventrally in a petri dish containing 1.5% low-melting-point agarose and visualized using a SZX16 fluorescence microscope (Olympus, Japan). The left optic nerve was transected proximally to its exit from the eye using sharpened insect pins. Partial transections were confirmed by the persistence of GFP-positive fibers. Injured larvae were released from the agarose and allowed to recover for 1 h in Ringer’s solution supplemented with 0.2 mM PTU, before being returned to E3 medium containing 0.2 mM PTU. At 24 h post-injury (hpi), lesion completeness was assessed by fluorescence: larvae with no detectable GFP signal in both the optic nerve and tectum were classified as complete transections and maintained in PTU-E3 medium for downstream fixation or live imaging. Larvae retaining GFP fluorescence in both the optic nerve and tectum were classified as partial transections.

For the transection experiments, each trial was conducted using 10 zebrafish larvae, from which 3 to 4 larvae with confirmed complete optic nerve transection were selected for further analysis.

As previously described [9], zebrafish larvae at 5 dpf were mounted dorsally in 1.5% low-melting agarose (prepared in E3 medium) in 35 mm glass bottom dishes (NEST Biotechnology, China). To prevent desiccation and ensure immobilization, larvae were maintained in E3 medium containing 0.2 mM PTU and 0.002% tricaine (Sigma-Aldrich, USA). Two-photon imaging was performed on transgenic zebrafish lines [Tg(isl2b.2:Gal4-VP16, myl7:EGFP); Tg(4XnrUAS:GFP)] using a multiphoton laser-scanning microscope (LSM 880 NLO with AiryScan; Carl Zeiss, Germany). GFP was excited using a Chameleon Ultra II IR laser (Coherent, USA), with the wavelength and intensity controlled using ZEISS (Zen Blue 3.1) software. The laser intensities were calibrated to 90, 85, and 86 mW for wavelengths of 800, 930, and 1,040 nm, respectively. Imaging was performed using an Objective W Plan-Apochromat 20 × (NA 1.05; Zeiss, Germany) at a resolution of 1,024 × 1,024 pixels. Z-plane adjustments were made at intervals of 2, 6, and 12 µm, and Zoom = 0.7 was applied as necessary to optimize the imaging depth of the volumetric structures. To enlarge the optic nerve region, a zoom factor of 1.5 was applied.

Further imaging was performed on double-transgenic zebrafish lines [Tg(Lyz:DsRed) and Tg(Isl2b.2:Gal4-VP16, myl7:EGFP); Tg(4XnrUAS:GFP)] using an FVMPE-RS multiphoton microscope (Olympus, Japan). GFP was excited using a Mai Tai HP DeepSee IR laser (Spectra-Physics, USA) at 920 nm (44.7 mW, 3% of 1.49 W total power), and DsRed was excited using an Insight X3 IR laser (Spectra-Physics, USA) at 1,100 nm (51.2 mW, 4% of 1.28 W total power). The imaging was performed using a 25 × water-immersion objective (NA 1.05; Olympus, Japan) at a resolution of 1,024 × 1,024 pixels. Z-plane adjustments were made at 2 µm intervals to optimize the imaging depth for visualizing the volumetric structures. All images were processed using Imaris software (Oxford Instruments, UK). Quantitative analyses of fluorescence intensity were performed using ImageJ (https://imagej.net/ij/). For the imaging experiments, each trial was performed using 3 to 6 zebrafish larvae.

OKR assays were performed as previously described [10]. Zebrafish larvae were collected by pipette with minimal water, briefly embedded in 1% low-melting-point agarose, and centrally positioned in 35 mm glass-bottom culture dishes. Their posture was then adjusted with fine forceps to avoid tissue injury. Upon agarose solidification, the culture dishes were carefully overlaid with E3 embryo medium. Subsequently, triangular apertures were precisely excised in the agarose to ensure unrestricted ocular motility. The preparations were then transferred to the OKR apparatus (ViewPoint, France). The stimulus parameters (i.e., contrast, color, spatial frequency, and angular velocity) were controlled using software. A 5-min training stimulus (100% contrast, black/white stripes) was provided before the experiments to stabilize tracking behaviors. Spatial frequencies were incrementally increased during testing, and eye movements were quantified by manual counting at 1-min intervals.

After the experiment, the larvae were gently extricated from the agarose and returned to the fresh system water. Eye rotations per minute were manually counted and analyzed using video recordings.

To determine the optimal stripe pattern for the OKR assay in zebrafish larvae, uninjured larvae were tested with grating frequencies of 0, 5, 10, and 20 stripes. A sample size of n = 6 larvae per group was used for each condition. In experiments involving PTU treatment, zebrafish larvae were exposed to PTU until 5 dpf. Subsequently, larvae were transferred to E3 medium and allowed to recover eye pigmentation over a period of 2.5 days. Following this recovery phase, OKR assays were conducted. Sample sizes were n = 9 larvae for the non-PTU control group and n = 10 for the PTU-treated group.

To minimize visual input from the contralateral eye, we conducted a monocular occlusion experiment. Healthy larvae were tested with monocular occlusion (n = 16) and without monocular occlusion (n = 16). In a subset of experiments, the uninjured eye was occluded with tin foil. For injured larvae, experiments were performed both with and without covering the uninjured eye. For the group without covering the uninjured eye, the sample sizes were as follows: uninjured (n = 9), 24 hpi (n = 6), and 72 hpi (n = 8). For larvae with the uninjured eye covered, the sample sizes were uninjured (n = 9), 24 hpi (n = 8), and 72 hpi (n = 8).

For enucleation experiments, viable larvae were selected for the OKR assay at 24 h post-enucleation. The sample sizes for OKR assessments were as follows: baseline (n = 7 for 8 dpf larvae, n = 12 for 5 dpf larvae), post-enucleation (n = 6 larvae), and post-contralateral optic nerve injury (n = 3 larvae).

All statistical analyses were performed using GraphPad Prism (version 9.5.0). One-way analysis of variance (ANOVA), followed by Dunnett’s test, was performed for comparisons among three or more groups. The unpaired t-test or Welch’s t-test was used for comparing data from two independent groups. Data are expressed as means ± standard error of the mean, and P < 0.05 were used to denote statistically significant differences.