Phenotype of newly generated neurons following spinal cord injury in zebrafish

Abstract

In contrast to humans, zebrafish can replace neurons that are lost to injury with new ones, which may contribute to structural and functional recovery after spinal cord injury. Despite this, the phenotypes, and the contribution of these newly regenerated neurons to functional recovery, remain unclear. To address this, I employed tamoxifen-activatable, genetic lineage tracing to specifically label cells which are newly generated from ependymo-radial glial (ERG) progenitors in response to injury. Here, temporal control of Cre-recombination made it possible to specifically label cells which were newly generated in response to spinal injury. Furthermore, to evaluate the phenotypes of regenerated neurons, I further optimized this technique, using Gateway cloning, which permitted specific assessment of lesion-generated neurons, their morphologies, and their morphological dynamics over time.

In this study, I first validated the molecular identity of recombined cells in the ERG-specific, genetic lineage tracing line, her4.3:CreERT2;β-actin:loxP-mCherry-loxP-EGFP in chapter 3. In larvae, I confirmed that GFP-positive cells which are labelled after Cre-recombination simultaneously expressed the pan-neuronal marker HuC, and incorporated the nucleotide analogue EdU, indicating that these cells are new neurons which proliferate following Cre-recombination. Furthermore, quantification of these co-labelled cells indicated that injury led to an 8-fold increase in neurogenesis, within 100 μm of the lesion site, compared to age-matched uninjured controls. In live observations in these animals, I saw ERGs dividing to generate new GFP-positive cells, between 10-17 hours post-lesion (hpl). These ERG-derived cells extended processes which subsequently crossed the lesion site by 48 hpl. Overall, this demonstrates that the cellular dynamics of newly generated cells can be observed in her4.3:CreERT2;β-actin:loxP-mCherry-loxP-EGFP animals. Furthermore, it validates that this tool reliably reports lesion-induced neurogenesis, following spinal cord injury in zebrafish larvae.

In Chapter 4, I assessed the phenotypes of new neurons, which are produced after injury. To do this, I introduced two changes to the β-actin:loxP-mCherry-loxP-EGFP loxP recombination reporter construct: Firstly, the neuron-specific promotor, NBT was used so that only neurons would be labelled, before and after Cre-recombination. Secondly, membrane-bound EGFP (EGFP-CAAX) was introduced so that morphology of neurons, which were labelled after Cre-recombination, could be visualised with a high degree of resolution. Mosaic expression of the optimised loxP recombination reporter construct (NBT:loxP-mCherry-loxP-EGFP-CAAX) in her4.3:CreERT2 animals, resulted in sparse labelling of newly generated neurons, which expressed EGFP-CAAX after Cre-recombination, and permitted quantification of individual neuronal morphologies. Following Cre-recombination in injured larvae, newly generated neurons expressed EGFP-CAAX were observed throughout the spinal cord, where 66% of them displayed morphologies which resembled those which were newly generated in uninjured animals (termed: developmentally appropriate), at 2 days post-lesion (dpl). At the same timepoint, a novel regeneration-specific phenotype was also observed, which was only generated within 100 μm of the injury site. Morphological assessment of the regeneration-specific phenotype at 2 dpl revealed that these neurons were significantly larger and more branched compared to new developmentally appropriate neurons in injured animals, or age-matched neurons in uninjured animals. By following the same neurons over time, it was also possible to assess growth dynamics between 2-3 dpl (5-6 days post-fertilisation). Regeneration-specific neurons showed accelerated neurite growth compared to both developmentally appropriate neurons in injured animals, and neurons in uninjured animals. Furthermore, regeneration-specific neurons were observed to rapidly arborise and extend neurites which successfully navigated into the distal injury site by 3 dpl, where they appeared to contact both pre-existing and new neurons suggesting possible integration into the spinal circuitry. The observation of an injury-generated neuronal phenotype which can extend neurites that cross the injury site following spinal cord injury, would resemble previously described “relay neurons” that are observed following spinal cord injury in adult zebrafish. Lastly, accelerated growth rates were found to be specific to regeneration-specific neuronal phenotype and were not observed in new developmentally appropriate neurons that were at equal distance to the lesion site. Developmentally appropriate neurons close to the lesion site also showed comparable growth rates to developmentally appropriate neurons far from the lesion site, as well as in uninjured animals. Together, these findings indicate that regeneration-specific neurons represent an injury-specific neuronal population which displays unique alterations to their intrinsic properties, which allows them to accelerate their morphological dynamics.

In summary, this study characterises a regeneration-specific neuronal phenotype which is generated by regenerative neurogenesis, following spinal cord injury in zebrafish. These regeneration-specific neurons show accelerated and irregular axon growth which is not observed in neurons with developmentally appropriate phenotypes. Furthermore, this thesis also establishes tissue-specific lineage tracing as a powerful tool to comprehensively evaluate the contribution of regenerative neurogenesis to functional recovery, following spinal cord injury.