Functional Regeneration of the Sensory Root via Axonal Invasion

Abstract

Regeneration following spinal root avulsion is broadly unsuccessful despite the regenerative capacity of other PNS-located nerves. By combining focal laser lesioning to model root avulsion in zebrafish, time-lapse imaging, and transgenesis, we identify that regenerating DRG neurons fail to recapitulate developmental paradigms of actin-based invasion after injury. We demonstrate that inducing actin reorganization into invasive components via pharmacological and genetic approaches in the regenerating axon can rescue sensory axon spinal cord entry. Cell-autonomous induction of invasion components using constitutively active Src induces DRG axon regeneration, suggesting an intrinsic mechanism can be activated to drive regeneration. Furthermore, analyses of neuronal activity and animal behavior show restoration of sensory circuit activity and behavior upon stimulating axons to re-enter the spinal cord via invasion. Altogether, our data identify induction of invasive components as sufficient for functional sensory root regeneration after injury.

Keywords: DREZ; DRG; OBPI; actin; behavior; invasion; regeneration; spinal cord; zebrafish.

Figure 1.. Taxol Rescues DRG Axon Spinal Entry after Avulsion-like Injury.

(A) Cross-section diagram of an intact and avulsed dorsal root. (B) Diagram of experimental model. At 3 dpf, a dorsal root is axotomized and time lapse imaged. (C) Z-projection time-lapse images of an avulsed DRG in a Tg(sox10:gal4); Tg(uas:lifeact-gfp) animal. Green arrows denote the growth cone. (D) Representative graph of axon length following injury. (E) Representative quantification of Lifeact-GFP intensity at the growth cone throughout regeneration. Red arrows represent actin structures, and brackets represent the duration of Lifeact peaks. (F) Representative scatterplot of the growth cone area and average Lifeact-GFP intensity during regeneration. Each time point is represented by a point. (G) Percentage of time points for which the regenerating growth cone displayed each actin organization. n = 5 DRG. (H) Outcomes of regeneration in 3 dpf injuries. n = 6 DRG per treatment. (I and J) Z-projection (I) and single-plane (J) images of Tg(ngn1:gfp); Tg(gfap:nsfb-mcherry) animals treated with DMSO or Taxol at 24 hpi. Injures at 3 and 5 dpf. Green arrows denote regenerated axons. (K) Outcomes of DRG injured at 3 or 5 dpf. n = 6 at 3 dpf, 7 at 5 dpf. (L) Outcomes of avulsed DRG treated with DMSO or Taxol at varying times after injury. n = 5 per treatment for 0–8 hpi, 6 per treatment for 12–24 hpi. Scale bar, 10 μm.

Figure 2.. Taxol Induces Actin-Rich Invasion Components during Spinal Re-entry.

(A–D) Z-projection time-lapse images of avulsed DRG in Tg(sox10:gal4); Tg(uas:lifeact-gfp) animals treated with DMSO (A), Taxol (B), Taxol+GM6001 (C), and Taxol+SU6656 (D). Graphs show Lifeact-GFP intensity tracings during regeneration. White boxes denote insets of the growing axon and growth cone regions. The green arrow denotes a stable invasive structure. Red arrows and brackets represent peaks and durations of peaks of Lifeact-GFP. (E) Graph of spinal re-entry in axons treated with DMSO, Taxol, Taxol+GM6001, and Taxol+SU6656. (F) Duration of actin concentrates in DMSO axons (n = 7), Taxol axons (n = 6), Taxol+GM6001 axons (n = 6), and Taxol+SU6656 axons (n = 4). (G) Representative scatterplot of growth cone area and average Lifeact-GFP intensity during regeneration in an axon treated with Taxol. Each time point is represented by a point. (H) Percentage of time points for which the regenerating growth cone displayed each actin organization. Tukey’s HSD was used in (D) and two-way ANOVA was used in (H). Scale bar, 10 mm.

Figure 3.. Induction of Invasion Components in DRG Neurons Promotes Regeneration.

(A–C) Z-projection time-lapse images of avulsed DRG in Tg(sox10:gal4); Tg(uas:lifeact-gfp) animals expressing uas:src-mcherry (A), uas:CA-src-mcherry (B), or uas:CA-src-mcherry and treated with Taxol (C). Graphs show Lifeact-GFP intensity tracings throughout regeneration. The green arrow denotes a stable invasive structure. Red arrows and brackets represent peaks of Lifeact-GFP. (D) Actin concentrate duration in Src DRG (n = 6), CA-Src DRG (n = 6), and CA-Src+Taxol DRG (n = 5). (E) Regeneration outcomes in DRG expressing sox10:Src, sox10:CA-Src, or sox10:CA-Src+Taxol or radial glial expressing gfap:CA-Src. (F) Percentage of time points the regenerating growth cone navigated using each actin organization. Tukey’s HSD was used in (D) and two-way ANOVA was used in (F). Scale bar, 10 μm.

Figure 4.. DRG Regeneration Restores Circuit and Behavioral Function.

(A) Z-projection time-lapse images of a Tg(sox10:syn-gfp) animal treated with Taxol before and after injury. White box denotes the area of insets. White arrows denote Syn-GFP puncta before injury. Green arrows denote new Syn-GFP puncta after re-entry. Blue arrows denote stabilized Syn-GFP. (B) Diagram of circuit and behavioral analysis. (C) Z-projection images of Tg(ngn1:gfp) animals stained for pErk following exposure to 4°C without injury or 8 consecutive avulsions and DMSO or Taxol treatment. Green outlines denote the area of DRG. Orange arrows denote pErk⁺ spinal cells. (D) Number of pErk⁺ cells in the spinal cord on ipsilateral and contralateral sides of injury (n = 8 animals per treatment). (E) Taxol-treated animals show a positive correlation between regenerated DRG and pErk intensity in the spinal cord (n = 8 animals). (F and G) Behavior at 24 hpi (F) and 48 hpi (G) in animals without injuries, avulsions with DMSO, and avulsions with Taxol (n = 8 at 24 hpi, 5 at 48 hpi). (H and I) Percentage of time shaking (H) and number of shaking behaviors (I) in animals without injuries, avulsions, and DMSO treatment and without avulsions and Taxol treatment 24 hpi (n = 8 animals per treatment). (J and K) Taxol-treated animals show a positive correlation between percentage of time shaking and regenerated DRG (J) and between percentage of time shaking and pErk spinal intensity (K) (n = 8 animals). Two-way ANOVA was used in (D); linear regression was used in (E), (J), and (K); and Tukey’s HSD was used in (H) and (I). Scale bar, 10 μm.