Fig 2: Different retinal ganglion cell (RGC) density in retinas of adult mice, zebrafish and killifish. (A–D) Scaled representation of retinas from young adult mice [10 weeks-old, (A)], young adult zebrafish [21 weeks-old, (B)], young adult killifish [6 weeks-old, (C)] and old killifish [18 week-old, (D)]. Mice have larger retinas than zebrafish and young killifish, while the retina of aged killifish is considerably larger than that of their younger counterparts. Scale bar 1 mm. (E–H) Representative micrographs of RGCs labeled with RBPMS [mouse, (E)] or Rbpms2 (fish), sampled from the temporal retina. Young adult zebrafish (F) and killifish (G) show a comparable density, higher than the one of old killifish (H) and young adult mice. Moreover, RGCs from the fish species are considerably smaller than the ones of mice. Scale bar 25 μm. (I) Automated quantification of the area of retinal WMs, revealing that unlike young adult fish, old killifish approach the size of murine retinas. (J) Automated quantification of RGC numbers in retinal WMs. Mice exhibit the lowest RGC count, with approximately 45,000 cells. In contrast, young fish possess around 70,000 RGCs. Aged killifish have the highest count, reaching approximately 125,000, nearly twice as many as young killifish. (K) Automated quantification of RGC density in retinal WMs. RGC density is considerably lower in mice compared to fish species. Notably, old killifish exhibit a significantly reduced RGC density compared to young adult fish. D, dorsal; N, nasal; RGCs, retinal ganglion cells; T, temporal; V, ventral; WMs, whole-mount.

Fig 3: Biphasic retinal ganglion cell (RGC) loss in killifish after optic nerve crush injury. (A) Representative image of a young adult (6 weeks-old) and old killifish (18 weeks-old) and experimental timeline for the RGC survival experiment, where RGC survival is evaluated at 4, 7, 14, and 21 days following ONC injury. Scale bar: 1 mm. (B) Representative micrographs of Rbpms2-stained WMs of young adult (6 weeks-old) and old killifish (18 weeks-old). For both ages, an appreciable loss of RGCs is evident at 7 dpi with further loss at 21 dpi, when compared to uninjured age-matched control fish. (C) Quantification of RGC survival in adult killifish WMs shows a first wave of RGC loss at 4 dpi, with 20% of the RGCs lost in both age groups, and this loss remains steady through 7 dpi. A second wave of loss is observed at 14 dpi, with older fish losing more RGCs (50%) compared to young fish (40%). No further loss is detected at 21 dpi in either age group. (D) Quantification of RGC survival per retinal quadrant in adult killifish WMs reveals no significant differences in inter-quadrant RGC loss after ONC by 21 dpi for both young adult (6 weeks-old) and old (18 weeks-old) fish. Data from two independent experiments, presented as percentages relative to the median of their uninjured age-matched control and presented as median ± 25–75th CI. Two-way Kruskal-Wallis ANOVA with pairwise Mann-Whitney U tests (C), One-way Kruskal-Wallis ANOVA and post hoc Mann-Whitney U test with Bonferroni correction [(D), 06 weeks], One-way Welch ANOVA and post hoc Games-Howell test [(D), 18 weeks]. p-values reported within the figure for significant differences. CI, confidence interval; DN, dorsonasal; dpi, days post injury; DT, dorsotemporal; ONC, optic nerve crush; RGCs, retinal ganglion cells; VN, ventronasal; VT, ventrotemporal; WMs, whole-mounts.

Fig 4: Minimal aberrant axonal growth and improved RGC survival after pONT in young adult killifish. (A) Representative biocytin-stained micrographs of retinal WMs of retrogradely traced young adult (6-week-old) killifish (N ≥ 5 per condition) showing very minimal aberrant RGC axon looping at the retinal margin from 10 dpi onwards (yellow arrowheads). Scale bar = 500 μm. (B) Representative micrographs of retinal WMs immunostained for Rbpms2 showing the dorsonasal (B) and ventrotemporal quadrants (C) from uninjured and injured (pONT) young adult killifish (N = ≥ 5 per condition). The uninjured retina displays regional RGC density differences, with higher RGC density in the ventrotemporal quadrant compared to dorsonasal. After pONT, a pronounced loss of RGCs is observed in the dorsonasal quadrant, corresponding to the injured optic nerve region, whereas the ventrotemporal quadrant, corresponding to the uninjured region, shows no evident cell loss. Scale bar = 50 μm. (C) Quantification of total RGC survival in Rbpms2-stained WMs shows a progressive decline starting from 4 dpi, retaining ~80% of RGCs at 65 dpi. Statistical analysis: two-way Kruskal–Wallis ANOVA followed by Mann–Whitney U tests with Bonferroni correction (N ≥ 5 per condition, presented as mean ± SEM). Significant differences (p < 0.05) are indicated within the graph. (D) Quadrant-based quantification reveals significant RGC loss in the DN quadrant by 10 dpi, with ~70% of dorsonasal RGCs remaining at 65 dpi. In contrast, the ventrotemporal quadrant shows minimal loss (~5% by 65 dpi). Statistical tests as in (D). Blue p-values denote differences between timepoints in the dorsonasal quadrant, and black p-values between quadrants at specific timepoints. AMC, age-matched controls; d, days; D, dorsal; DN, dorsonasal; dpi, days post injury; DT, dorsotemporal; N, nasal; pONT, partial optic nerve transection; RGC, retinal ganglion cell; V, ventral; VN, ventronasal; VT, ventrotemporal; SEM, standard error of the mean; T, temporal, WM, whole-mount.

Fig 5: Age-dependent difference in retinal ganglion cell survival and immune response after cONT. (A,B) Representative Rbpms2-stained micrographs of retinal whole-mounts from young adult [6-week-old, (A)] and aged [18-week-old, (B)] killifish (N ≥ 5 per condition) with and without injury (cONT). Both age groups demonstrate marked RGC loss at 21 and 65 dpi compared to their uninjured AMCs, retaining similar low RGC densities after cONT. At 10 dpi, aged killifish display visibly higher RGC densities than young adults. Scale bar = 50 μm. (C) Quantification of RGC survival in adult killifish WMs after cONT reveals progressive loss in young adults starting from 4 dpi, while old fish show a milder decline during the first 10 dpi. By 65 dpi, both groups converge to ~25% RGC survival. Statistical analysis: two-way Kruskal–Wallis ANOVA and post hoc Mann–Whitney U tests with Bonferroni correction (N ≥ 5 per condition, mean ± SEM). Orange, red, and black p-values indicate differences between timepoints in young adults, aged fish, and between age groups at specific timepoints, respectively. (D) Representative Lcp1-stained (pink) midsagittal retinal sections from young adult and aged killifish after cONT (N = 4 per condition). Old killifish display a higher microglia/macrophage density in the uninjured retina compared to their younger counterparts. Young adults show a peak in immune activation at 4 dpi, whereas aged fish exhibit a delayed neuroinflammatory response only peaking at 21 dpi. The analyzed region of interest is delineated by a white dotted line, and individual microglia/macrophages are indicated by yellow arrowheads. Scale bar = 50 μm. (E) Quantification of retinal microglia/macrophage density showing higher baseline levels in aged compared to young adult killifish. Following cONT, immune cell density in young adult fish rapidly increases, peaking at 4 dpi, and surpasses that of aged fish. By 21 dpi, densities return to baseline in young adults, whereas aged fish show maximal accumulation. Statistical analysis as in (C) (N = 4 per condition, mean ± SEM). Orange, red, and black p-values correspond to intra- and inter-age group comparisons as indicated in (C). (F) Normalized retinal microglia/macrophage densities relative to uninjured AMC. In young adults, immune activation peaks at 4 dpi with a fourfold increase, while aged fish reach maximal densities at 21 dpi, nearly doubling relative to baseline. Statistical analysis: ordinary one-way ANOVA and post-hoc Turkey’s multiple comparison test (N = 4 per condition, mean ± SEM). Orange and red p-values indicate intra-age comparisons to the uninjured condition for young and aged fish, respectively. AMC, age-matched controls; cONT, complete optic nerve transection; d, days; DAPI, 4′,6-diamidino-2-phenylindole; dpi, days post injury; GCL, ganglion cell layer; IPL, inner plexiform layer; INL, inner nuclear layer; Lcp1, lymphocyte cytosolic protein 1; OPL, outer plexiform layer; ONL, outer nuclear layer; Rbpms2, RNA binding protein with multiple splicing 2; RGC, retinal ganglion cell; SEM, standard error of the mean; Uninj., uninjured.