Glycine-alanine dipeptide repeat protein contributes to toxicity in a zebrafish model of C9orf72 associated neurodegeneration

Abstract

Background: The most frequent genetic cause of frontotemporal lobar degeneration (FTLD) and amyotrophic lateral sclerosis (ALS) is the expansion of a GGGGCC hexanucleotide repeat in a non-coding region of the chromosome 9 open reading frame 72 (C9orf72) locus. The pathological hallmarks observed in C9orf72 repeat expansion carriers are the formation of RNA foci and deposition of dipeptide repeat (DPR) proteins derived from repeat associated non-ATG (RAN) translation. Currently, it is unclear whether formation of RNA foci, DPR translation products, or partial loss of C9orf72 predominantly drive neurotoxicity in vivo. By using a transgenic approach in zebrafish we address if the most frequently found DPR in human ALS/FTLD brain, the poly-Gly-Ala (poly-GA) protein, is toxic in vivo.

Method: We generated several transgenic UAS responder lines that express either 80 repeats of GGGGCC alone, or together with a translation initiation ATG codon forcing the translation of GA80-GFP protein upon crossing to a Gal4 driver. The GGGGCC repeat and GA80 were fused to green fluorescent protein (GFP) lacking a start codon to monitor protein translation by GFP fluorescence.

Results: Zebrafish transgenic for the GGGGCC repeat lacking an ATG codon showed very mild toxicity in the absence of poly-GA. However, strong toxicity was induced upon ATG initiated expression of poly-GA, which was rescued by injection of an antisense morpholino interfering with start codon dependent poly-GA translation. This morpholino only interferes with GA80-GFP translation without affecting repeat transcription, indicating that the toxicity is derived from GA80-GFP.

Conclusion: These novel transgenic C9orf72 associated repeat zebrafish models demonstrate poly-GA toxicity in zebrafish. Reduction of poly-GA protein rescues toxicity validating this therapeutic approach to treat C9orf72 repeat expansion carriers. These novel animal models provide a valuable tool for drug discovery to reduce DPR associated toxicity in ALS/FTLD patients with C9orf72 repeat expansions.

Keywords: C9orf72; Zebrafish; poly-GA toxicity.

Fig. 1

Generation of transgenic zebrafish model of C9ORF72 repeat expansion disease. a Schematic representation of Gal4 driver line zebrafish crossed to a UAS responder transgenic zebrafish to generate embryos that express a transgene under the control of the UAS. Schematic representation of the responder constructs used for the generation of transgenic zebrafish. b Genotyping by PCR of 1 dpf embryos. pCS2 + 2xGGGGGCC and pCS2 + 80XGGGGCC constructs were used as standards for GA2-GFP and GA80-GFP in lane 2 and 3. Positions of 2xGGGGCC and 80xGGGGCC repeats are indicated by arrows. c Semi-quantitative RT-PCR analyses for the wild-type, GA2-GFPa/b, GA80-GFPa/b and ggggcc80-GFP zebrafish. Note that all transgenic lines showed similar expression level at 4 dpf embryos. d Immunoblotting of wild-type, GA2-GFPa/b, GA80-GFPa/b and ggggcc80-GFP with antibodies as indicated with embryonic lysates of 4 dpf old embryos

Fig. 2

RNA foci formation in transgenic zebrafish. a, b Cy3-labeled in situ probe detected dot-like structures in spinal cord in GA80-GFPa/b and ggggcc80-GFP zebrafish. b Foci were only detected in GA80-GFP and ggggcc-GFP fish whereas no foci were detected in wild-type and GA2-GFPa/b fish at 28 hpf. GA80-GFPa zebrafish were treated with RNaseA or DNase. Scale bar 10 μm. c Pericardial edema phenotype observed in ggggcc80-GFP zebrafish at 4 dpf. Phenotypic features are classified as wild-type, mild edema, and severe edema. d The average percentages of phenotypic fish of the three different phenotypic classes at 4 dpf are indicated in the bargraph (at least three independent clutches were analyzed with n ≥ 14)

Fig. 3

Poly-GA protein elicits a toxic phenotype. a In vivo image of GA-GFP polypeptides at 2 and 4 dpf. Genotypes as indicated. No GFP sibling (GA80-GFP) refers to a sibling from a cross between a Gal4 driver and a UAS GA80-GFP responder fish that is GFP negative, and hence is either negative for the driver or the responder construct, or both constructs. GFP fluorescent images shown are merged with DIC pictures. Lowest panel is a magnification of the middle panel at 4 dpf. Lateral views of the trunk musculature. Scale bar 20 μm. b A strong pericardial edema phenotype was observed in GA80-GFPa/b zebrafish at 4 dpf. The average percentages of phenotypic fish of the three different classes are indicated in the bargraph. c GA80-GFPa/b zebrafish had mostly no circulation at 2 dpf. Red blood cells accumulate due to circulation defects (arrow). d The average percentages of fish with or without circulation are indicated in the bargraph (at least three independent clutches were analyzed with n ≥14)

Fig. 4

Antisense morpholino rescued the toxic edema phenotype. a Representation of process of intervention for each morpholino during the generation of GA80-GFP protein in vivo. b GAL4 targeting AMO efficiently blocked Gal4 translation at 2 dpf shown by immunoblotting (upper panel). Quantification of the pericardial edema phenotype observed in GA80-GFP with injection of ctrl AMO or GAL4 targeting AMO at 4 dpf are shown as a bar graph (lower panel) (p < 0,001, 3 independent experiments with 3 clutches n ≥ 6 are shown, unpaired t test). c ATG targeting morpholino efficiently inhibited the ATG dependent translation of poly-GA at 2 dpf (upper panel). Quantification of the pericardial edema phenotype observed in GA80-GFP upon injection of ctrl AMO or ATG targeting AMO at 4 dpf shown as a bar graph (lower panel) (p < 0,005, 3 independent experiments with 3 clutches n ≥ 19 are shown, unpaired t test). d Semi-quantitative RT-PCR analyses of injected embryos at 2 dpf. e RNA foci formation was not affected upon injection with ctrl AMO or ATG AMO at 2 dpf. Scale bar 10 μm