A Critical Review of Zebrafish Neurological Disease Models-1. The Premise: Neuroanatomical, Cellular and Genetic Homology and Experimental Tractability

Abstract

The last decade has seen a dramatic rise in the number of genes linked to neurological disorders, necessitating new models to explore underlying mechanisms and to test potential therapies. Over a similar period, many laboratories adopted zebrafish as a tractable model for studying brain development, defining neural circuits and performing chemical screens. Here we discuss strengths and limitations of using the zebrafish system to model neurological disorders. The underlying premise for many disease models is the high degree of homology between human and zebrafish genes, coupled with the conserved vertebrate Bauplan and repertoire of neurochemical signaling molecules. Yet, we caution that important evolutionary divergences often limit the extent to which human symptoms can be modeled meaningfully in zebrafish. We outline advances in genetic technologies that allow human mutations to be reproduced faithfully in zebrafish. Together with methods that visualize the development and function of neuronal pathways at the single cell level, there is now an unprecedented opportunity to understand how disease-associated genetic changes disrupt neural circuits, a level of analysis that is ideally suited to uncovering pathogenic changes in human brain disorders.

Keywords: CRISPR; disease models; neural development; neuroanatomy; transgenics; zebrafish.

Figure 1

Fluorescence microscopy and photoconversion in live transgenic zebrafish. Confocal Z-plane projections of zebrafish larva expressing the photoconvertible fluorophore Kaede in a pan-neuronal pattern throughout the CNS at 3 dpf. At baseline, fluorescence is seen in the green, but not red, channel (above). Immediately following exposure to UV light, fluorescence is seen only in the red channel (below)

Figure 2

Imaging and manipulating RedOx biochemistry in vivo. (A) Ratiometric intravital confocal images of diencephalic neurons expressing the glutathione oxidation reporter GRX1-roGFP2. Dramatic changes in the fluorescence excitation properties of the reporter following application of hydrogen peroxide to the bath reveal a significant change in the intracellular glutathione RedOx potential [35]. (B) Intravital confocal images of mCerulean3-labeled mitochondria in the lateral line nerve of transgenic zebrafish also expressing a mitochondrially targeted form of the fluorogen activating protein dL5^**. Exposure to far-red light in the presence (below), but not absence (above), of the iodine substituted fluorogen MG2I results in mitochondrial fragmentation as a consequence of singlet oxygen formation within the mitochondrial matrix [36].

Figure 3

Comparative neuroanatomy of human and larval zebrafish. Schematic representations of sagittal sections of adult human (above) and 5-day post-fertilization zebrafish (below) brains. Homologous areas are shaded the same colors to illustrate the conservation of basic vertebrate brain organization from human to larval zebrafish, despite the dramatic difference in scale and complexity.

Figure 4

Conservation of protein coding genes in zebrafish and mouse models. Human protein-coding genes with orthologs in zebrafish (left) and mouse (right) shaded in purple, while those without orthologs are shaded red. Blue shading for zebrafish and mouse specific genes. The subset of human genes known to cause disease is also indicated. Data downloaded from ensembl biomart, filtering for protein-coding genes (GRCh38.p13). Genes on non-main contigs (i.e. patches and alternative assemblies) are excluded. Disease-associated genes were downloaded using the MIM morbid description annotation.

Figure 5

A high fraction of neurodevelopmental disease genes have zebrafish orthologs. Chart indicates the percent of human genes with zebrafish paralogs, for all 19 988 human protein coding genes (left), and the 1052 human disease genes that are highly expressed during brain development (right). Outer rings show the percentage of human genes with either no (grey) or at least one zebrafish paralog (blue), while inner rings break down zebrafish genes according to whether zebrafish have one, two or more than three paralogs. Human protein-coding genes and their zebrafish homologs were downloaded from ensembl biomart, using the MIM morbid description to designate disease genes. For genes highly expressed during human neural development we downloaded the BrainSpan developmental transcriptome dataset (RNA-Seq Gencode v10 archive) [164] and identifed those genes whose maximal expression across developmental stages and timepoints were in the top 10% quantile.

Figure 6

Neuropeptide conservation in zebrafish. Number of paralogs of each major human neuropeptide in zebrafish. Green bar: Human genes that lack a zebrafish paralog. Successive bars: Human genes with 1 to 4 zebrafish paralogs each. The list of neuropeptides was taken from the neuropeptide database [89], and zebrafish homologs identified for each using ZFIN, PANTHER phylogenetic trees and reciprocal blast [165, 166]

Figure 7

Spinal cord regeneration in a larval zebrafish. Serial confocal Z-plane projections show the spinal cord of a zebrafish (A) at 10 dpf before any experimental manipulation, (B) at the same age immediately after spinal cord transection, and (C) 2 days later. Axons of projection neurons are labeled with membrane bound EGFP. Disrupted spinal tracts in (B) have regenerated in (C), such that abundant axons are seen crossing the transection site at the later time point. Lesion and imaging performed as previously described [167].