Serotonin neuromodulation directs optic nerve regeneration

Abstract

Optic nerve (ON) regeneration in mammalian systems is limited by an overshadowing dominance of inhibitory factors. This has severely hampered the identification of pro-regenerative pathways. Here, we take advantage of the regenerative capacity of larval zebrafish to identify pathways that promote ON regeneration. From a small molecule screen, we identified modulators of serotonin (5-HT) signaling that inhibit ON regeneration. We find that several serotonin type-1 (5-HT1) receptor genes are expressed in retinal ganglion cells during regeneration and that inhibiting 5-HT1 receptors or components of the 5-HT pathway selectively impedes ON regeneration. We show that 5-HT1 receptor signaling is dispensable during ON development yet is required for regenerating axons to emerge from the injury site. Blocking 5-HT receptors once ON axons have crossed the chiasm does not inhibit regeneration, suggesting a selective role for 5-HT receptor signaling early during ON regeneration. Finally, we show that agonist-mediated activation of 5-HT1 receptors leads to enhanced and ectopic axonal regrowth. Combined, our results provide evidence for mechanisms through which serotonin-dependent neuromodulation directs ON regeneration in vivo.

Keywords: 5HT; Axon regeneration; Optic nerve; Serotonin; Zebrafish.

Fig. 1.

Regenerating RGC axons robustly regrow to the optic tecta by 48 h post-transection. (A-E) Fluorescent representative images of Tg(Isl2b:GFP) fixed larvae visualized using confocal imaging at distinct time points: pre-transection, 24 h post-transection, 32 h post-transection and 48 h post-transection. Arrows depict the midline at the optic chiasm. Different larvae were used across time points. Scale bars: 50 μm. (A,A′) At 5 dpf, RGC axons have crossed the midline at the optic chiasm (top) and innervated the contralateral tecta (bottom). (B) Timeline of optic nerve transection, regeneration and larvae exposure to small molecules. Schematics describe the pre-transection and post-transection (0, 24, 32 and 48 hpt) time points during optic nerve regeneration and its assessment in larval zebrafish. (B′) At 0 hpt, the left optic nerve of Tg(Isl2b:GFP) larvae was transected, and the right eye was removed as detailed in the Materials and Methods. (C,C′) At 24 hpt, transected axons begin to extend towards the optic chiasm (chiasm region; C, top). Complete transection of the optic nerve was corroborated by examining larvae dorsally (tecta region; C, bottom) at 22-24 hpt for remnants of Isl2b:GFP nerve structures. Degenerated tecta are outlined with dashed lines. (D,D′) At 32 hpt, a small group of regeneration RGC axons have reached the midline (D, top). Degenerated tecta are outlined with dashed lines (D, bottom). (E,E′) At 48 hpt, wild-type larvae regrow their RGC axons toward the optic chiasm (E, top) and begin re-innervating the peripheral edges of the tecta (E, bottom). Live zebrafish larvae treated with small molecules at 24 hpt were assessed to determine whether optic nerve regeneration was impaired. Top and bottom panels for each time point are of the same larva. At 48 hpt, the ipsilateral tectum (left tectum) is typically more innervated than the contralateral tectum. Images are representative of at least six different samples per time point.

Fig. 2.

5-HT1 receptor signaling promotes early optic nerve regrowth without impairing overall CNS axonal regrowth. (A-C′) Fluorescent representative images (A-C) and schematics (A′-C′) of Tg(Isl2b:GFP) at 48 hpt of larvae exposed to 0.3% DMSO (A) and 50 μM of antagonist WAY-100635 (B,C). Images in B and C are examples of two different types of regrowth observed when applying this small molecule. White arrows indicate the midline region at the optic chiasm. Scale bar: 50 μm. (B) RGC axons of larvae treated with the antagonist WAY-100635 regrew toward the optic chiasm (top, yellow arrowhead) but failed to innervate the contralateral tectum (dashed lines, bottom). A group of axons regrew ectopically away from the optic chiasm (yellow arrow). Re-innervation of the ipsilateral tectum begins soon after the small molecules are added. (C) Regenerating RGC axons of a larva treated with the antagonist WAY-100635 stalled near the injury site (yellow arrowhead, top). No axonal regrowth is observed in the optic chiasm or contralateral tectum (dashed lines, bottom). (D) Quantification of optic nerve axonal re-innervation to contralateral tectum at 48 hpt in Tg(Isl2b:GFP) larvae treated from 24 hpt to 48 hpt with 0.3% DMSO or WAY-100635. Bars represent the relative rate of optic nerves that fail to re-innervate the contralateral tectum in WAY-100635-treated (gray bars) and control (white bars) groups, plotted as a log₂ scale. See Materials and Methods for details on ratios and fold-change calculations. *P<0.05 (two-tailed Fisher's exact test). ns, not significant. n=21 and n=25 for nerves treated with DMSO and 5 μM WAY-100635, respectively; n=37 and n=45 for nerves treated with DMSO and 50 μM WAY-100635, respectively. (E) Quantification of ectopic regrowth at 48 hpt in Tg(Isl2b:GFP) larvae treated from 24 hpt to 48 hpt with 0.3% DMSO or WAY-100635. Bars represent the relative rate of optic nerves' ectopic regrowth in WAY-100635-treated (black bars) and control (white bars) groups, plotted as a log₂ scale. See Materials and Methods for details on ratios and fold-change calculations. The data displayed in D,E were obtained from larvae in the same treated groups. Statistics were determined using the two-tailed Fisher's exact test; ns, not significant. n=21 and n=25 for nerves treated with DMSO and 5 μM WAY-100635, respectively; n=37 and n=45 for nerves treated with DMSO and 50 μM WAY-100635, respectively. (F) Top: Timeline of optic nerve regeneration from 0 hpt-48 hpt. DMSO or small molecules were added to Tg(Isl2b:GFP) larvae at 32 hpt, and optic nerve regeneration was assessed at 48 hpt. Larvae were kept in 1× PTU/E3 from 24 hpt to 32 hpt. Bottom: Quantification of optic nerve axonal re-innervation to contralateral tecta at 48 hpt in Tg(Isl2b:GFP) larvae treated with 0.3% DMSO (white bars) or 50 μM WAY-100635 at 24 and 32 hpt (gray bars). Bar graphs and the ratios observed were calculated as detailed in D and Materials and Methods. The data showing 50 μM WAY-100635 at 24 hpt is identical to that shown in D and is shown here for visual comparison only. *P<0.05 (two-tailed Fisher's exact test); ns, not significant. n=37 and n=45, for nerves treated with DMSO and WAY-100635 at 24 hpt, respectively. n=13 and n=38, for nerves treated with DMSO and WAY-100635 at 32 hpt, respectively. (G) Top: Timeline of Mauthner axon regeneration. Schematic shows a laser transection (orange laser) performed in Tg(Tol-056:GFP) 5 dpf larvae using a UV laser (355-nm wavelength) at the ninth spinal cord hemisegment. After transection, 0.3% DMSO or 50 μM WAY-100635 was added to larvae starting at 1 hpt (dropper). Fresh DMSO and small molecules were replenished at 24, 48 and 72 hpt. Mauthner axon regrowth was assessed at 96 hpt. Bottom: Quantification of axonal length regrowth at 96 hpt, measured in spinal cord segments in control and WAY-100635 treated groups. Statistics were determined using the one-way ANOVA test. ns, not significant. Error bars represent s.e.m.

Fig. 3.

5-HT1 receptor genes are expressed in RGCs at pre-transection and during optic nerve regeneration. (A-F′) Representative images of retinas from Tg(isl2b:GFP) 5 dpf uninjured larvae stained with no probe control (A-C; n=8 retinas) or an in situ hybridization HCR probe mix for htr1aa, htr1ab, htr1b and htr1d (D-F′; n=12) (magenta). Images shown are maximum z-projections of six horizontal optical sections (32 μm). Dashed white boxes in E,F highlight the area enlarged F′, a merged maximum z-projection of ten horizontal optical sections (0.1 μm) showing multiple Tg(isl2b:GFP) RGC neurons (green, outlined with yellow dashed line) expressing mRNA of 5-HT1 receptors (magenta). In F′, the brightness of the green channel was adjusted for better visualization of 5-HT1 receptor genes inside RGC neurons. White asterisks depict nonspecific staining. The retinal pigmented epithelium (pe) was nonspecifically labeled by the amplifier with Alexa Fluor 546. (G-L′) Representative images of retinas from Tg(isl2b:GFP) 5 dpf uninjured larvae (G-I′; n=7) or larvae with transected optic nerves (J-L′; n=5) at 24 hpt stained with an in situ hybridization HCR probe mix for htr1aa and htr1ab (magenta). Images shown are maximum z-projections of seven horizontal optical sections (G-I) and 12 optical sections (J-L) (32 μm). Dashed white boxes in H,I highlight the area enlarged in I′, a merged maximum z-projection of 12 horizontal optical sections (0.1 μm). In I′, yellow dashed lines outline cell bodies with mRNA expression. The brightness of the green channel was adjusted for better visualization of 5-HT1 receptor genes inside RGC neurons. Dashed white boxes in K,L highlight the area enlarged in L′, a merged maximum z-projection of 11 horizontal optical sections (0.1 μm). In L′, yellow dashed lines outline cell bodies with mRNA expression. The brightness of the green channel was adjusted for better visualization of 5-HT1 receptor genes inside RGC neurons. White asterisks depict nonspecific staining. The retinal pigmented epithelium (pe) was nonspecifically labeled by the amplifier with Alexa Fluor 546. Scale bar: 50 μm (A-C,D-F,G-I,J-L).

Fig. 4.

Optic nerve development is unaffected after treating zebrafish embryos with 5-HT1A antagonists. (A) Timeline of optic nerve development from 24-48 hpf and larvae exposure to small molecules. DMSO (0.3%) or 50 μM of antagonist WAY-100635 was added to embryos at 24 hpf. Around 32 hpf, RGCs differentiate and immediately begin extending towards the brain. At 48 hpf, RGC axons have extended past the optic chiasm and have begun innervating the optic tecta. At this time point, treatment was washed off and optic nerve development was assessed. Developmental timeline information taken from Chhetri et al. (2014). (B,C) Representative images of 48 hpf Tg(Isl2b:GFP) embryos treated with 0.3% DMSO (B) and 50 μM WAY-100635 (C). White arrows depict the midline at the optic chiasm. No apparent growth or guidance defects are observed in the developing optic nerves of control and small molecule-treated larvae. Scale bar: 50 μm. (D) Quantification of optic nerve growth defects at 48 hpf in Tg(Isl2b:GFP) larvae treated from 24 hpf to 48 hpf with DMSO (white bar) or WAY-100635 (black bar). Bars represent the percentage of defective optic nerves for each treatment group. Statistics were determined using a two-tailed Fisher's exact test. ns, not significant. n=46 and n=33, for larvae treated with DMSO and WAY-100635, respectively.

Fig. 5.

Exogenous activation of 5-HT1 receptors significantly increases axonal regrowth in a dose-dependent manner. (A-B′) Fluorescent representative images and schematics of Tg(Isl2b:GFP) at 48 hpt of larvae exposed to 0.3% DMSO (A) and 50 μM of the agonist zolmitriptan (B,B′). Images in B and B′ are examples of two different types of regrowth observed when applying this small molecule. Scale bar: 50 μm. (B) Regenerating optic nerve of a larva treated with 50 μM zolmitriptan shows axons exhibiting ectopic regrowth at the optic chiasm (white arrowhead). RGC axons also regrew misguided away from the chiasm to a more posterior region of the brain (yellow arrowhead). Other RGC axons re-innervated both optic tecta by 48 hpt (bottom image). (B′) The regenerating optic nerve of a larva treated with 50 μM zolmitriptan shows a significant amount of axon fascicles ectopically regrowing from the injury site and extending away from the optic chiasm (yellow arrowheads). RGC axons did not re-innervate the contralateral tectum by the 48 hpt time point (bottom image, dashed lines). White arrows depict the midline at the optic chiasm. (C) Quantification of optic nerve axonal re-innervation to the contralateral tectum at 48 hpt in larvae treated from 24 hpt to 48 hpt with 0.3% DMSO or the agonist zolmitriptan. Bars represent the relative rate of optic nerves that re-innervated the contralateral tectum in zolmitriptan-treated (gray bars) and control (white bars) groups, plotted as a log₂ scale. See Materials and Methods for details on ratios and fold-change calculations. The data displayed in C,D were obtained from larvae in the same treated groups. *P<0.05 (two-tailed Fisher's exact test); ns, not significant. n=19 and n=23 for nerves treated with DMSO and 0.5 μM zolmitriptan, respectively; n=27 and n=33 for nerves treated with DMSO and 5 μM zolmitriptan, respectively; and n=22 and n=27 for nerves treated with DMSO and 50 μM Zolmitriptan, respectively. (D) Quantification of ectopic regrowth at 48 hpt in Tg(Isl2b:GFP) larvae treated from 24 hpt to 48 hpt with 0.3% DMSO or zolmitriptan. Bars represent the relative rate of optic nerves' ectopic regrowth in zolmitriptan-treated (black bars) and control (white bars) groups, plotted as a log₂ scale. See Materials and Methods for details on ratios and fold-change calculations. *P<0.05 (two-tailed Fisher's exact test); ns, not significant. n=19 and n=23 for nerves treated with DMSO and 0.5 μM zolmitriptan, respectively; n=27 and n=33 for nerves treated with DMSO and 5 μM zolmitriptan, respectively; and n=22 and n=27 for nerves treated with DMSO and 50 μM zolmitriptan, respectively.