Serotonin Promotes Development and Regeneration of Spinal Motor Neurons in Zebrafish

Abstract

In contrast to mammals, zebrafish regenerate spinal motor neurons. During regeneration, developmental signals are re-deployed. Here, we show that, during development, diffuse serotonin promotes spinal motor neuron generation from pMN progenitor cells, leaving interneuron numbers unchanged. Pharmacological manipulations and receptor knockdown indicate that serotonin acts at least in part via 5-HT1A receptors. In adults, serotonin is supplied to the spinal cord mainly (90%) by descending axons from the brain. After a spinal lesion, serotonergic axons degenerate caudal to the lesion but sprout rostral to it. Toxin-mediated ablation of serotonergic axons also rostral to the lesion impaired regeneration of motor neurons only there. Conversely, intraperitoneal serotonin injections doubled numbers of new motor neurons and proliferating pMN-like progenitors caudal to the lesion. Regeneration of spinal-intrinsic serotonergic interneurons was unaltered by these manipulations. Hence, serotonin selectively promotes the development and adult regeneration of motor neurons in zebrafish.

Graphical abstract

Figure 1

Serotonin Signaling Promotes Embryonic Motor Neuron Generation Lateral views of spinal cords at 33 hpf are shown. (A–F) Serotonin (5-HT) treatment (24–33 hpf) increases the number of HB9 immuno-labeled motor neurons but has no influence on vsx1:GFP (A–C) and pax2a:GFP labeled interneurons (D–F) in the same embryos (Student’s t test in C, ^∗∗p = 0.0077; in F, ^∗∗p = 0.002). (G–I) Serotonin treatment increases the number of dividing (pH3⁺) olig2:GFP⁺ pMN progenitor cells (Student’s t test in I; ^∗∗∗p = 0.0006). (J–L) Lateral view of a double-transgenic olig2:dsRed/HB9:GFP embryo is shown with red only (arrows, pMN progenitors) and double-labeled (arrowheads, motor neurons) cells indicated in the spinal cord (J). A typical FACS profile is shown (K). In RT-PCR, serotonin receptors show enrichment in pMN progenitor cells, compared to motor neurons (L). GAPDH is used for comparison. (M–O) Morpholino knockdown of receptor htr1ab reduces the number of HB9⁺ motor neurons but does not influence the number of vsx1:GFP⁺ interneurons in the same embryos (Student’s t test; ^∗∗∗p < 0.0001). The scale bar in (B) represents 10 μm for (A) and (B), in (E) represents 10 μm for (D) and (E), in (H) represents 15 μm for (G) and (H), and in (N) represents 15 μm for (M) and (N). See also Figure S1.

Figure 2

Serotonergic Axons Are Located in Close Proximity to Processes of pMN-like Adult Progenitor Cells Only Rostral to the Lesion Spinal cross-sections are shown (dorsal is up; asterisk indicates central canal). (A and B) Timeline and schematic representations of serotonergic axons in the injured spinal cord in lateral (A) and cross-sectional views (B), indicating the areas of analysis and density of serotonergic axons in relation to pMN-like ERGs. (C–F) Overviews of 5-HT⁺ labeling in olig2:GFP fish rostral (C and C’) and caudal (E and E’) to the lesion are shown. In the ventro-lateral area, close apposition of serotonergic axons (red) with radial processes (green, arrows) of olig2:GFP⁺ pMN-like progenitor cells (arrowheads) are visible rostral (D), but not caudal (F), to the lesion. (G and G’) Olig2:dsRed⁺/mbp:GFP⁻ ERGs (arrows), olig2:dsRed⁺/mbp:GFP⁺ oligodendrocytes (arrowheads), and myelin sheaths (open arrowheads) are indicated (G, overview; G’, detail). (H) PCR after FACS indicates dsRed-only-labeled cells are enriched in the pMN-like ERG fraction, whereas GFP is only amplified in the oligodendrocyte fraction (OLs). EF1alpha is included for comparison. hrt1aa and hrt3a are expressed in adult pMN-like ERGs. The scale bar in (E’) represents 50 μm for (C), (C’), (E), and (E’); in (F) represents 25 μm for (D) and (F); in (G) represents 25 μm; and in (G’) represents 10 μm. See also Figure S2.

Figure 3

Serotonergic, but Not Dopaminergic, Descending Axons Are Ablated by 5,7-DHT Spinal cross-sections are shown (dorsal is up; asterisk indicates central canal). (A–C) An intraperitoneal injection of 5,7-DHT ablates serotonergic (5-HT), but not TH1⁺, axons within 2 days (Mann-Whitney U-test; ^∗p = 0.0357). (D–I) At 14 dpl, the number of axons rostral to the lesion is still strongly reduced after prior ablation (D–F), whereas caudal to the lesion, the low axon density is unaltered (G–I; Mann-Whitney U-test; ^∗∗p = 0.0022). The scale bar in (H) represents 50 μm. See also Figure S3.

Figure 4

Ablation of Serotonergic Axons Inhibits Regeneration of Motor Neurons, but Not Serotonergic Neurons Spinal cross-sections are shown (dorsal is up; asterisk indicates central canal). (A) In the unlesioned spinal cord, few small HB9⁺ and motor neurons and 5-HT⁺ cells (arrow) are present. (B–G) Rostral to the lesion (see timeline for experimental condition), ablation of serotonergic axons leads to reduced motor neuron (arrowheads) regeneration without influencing regeneration of serotonergic neurons (arrows). Caudal to the lesion, numbers of newly generated motor neurons and serotonergic neurons are unaltered (Mann-Whitney U-test; ^∗p = 0.0344). The scale bar in (E) represents 25 μm. See also Figure S4.

Figure 5

Serotonin Injections Increase the Number of Newly Generated Motor Neurons and Proliferating pMN-like Progenitor Cells Caudal to the Spinal Lesion Cross-sections through the spinal cord are shown; asterisks indicate the central canal, arrowheads indicate PCNA⁺/olig2:GFP⁺ cells, solid arrows indicate EdU⁺/olig2:GFP⁺ cells, and empty arrows indicate PCNA⁺/olig2:GFP⁻ or EdU⁺/olig2:GFP⁻ cells. (A–E) Serotonin injection doubles the number of newly generated motor neurons caudal to the lesion but has no effect rostral to the lesion (see timeline in E for experimental paradigm; Student’s t test; ^∗∗∗p < 0.0001). (F–I) In the ventricular zone of olig2:GFP transgenic animals, the numbers of PCNA⁺/olig2:GFP⁺ (Mann-Whitney U-test; ^∗p = 0.0437) and EdU⁺/olig2:GFP⁺ (Mann-Whitney U-test; ^∗p = 0.0231) pMN-like ERGs are significantly increased only caudal to the lesion (see timeline for experimental paradigm). The scale bar in (D) represents 50 μm for (A)–(D) and in (G’’’) represents 10 μm for (F) and (G). See also Figure S5.