This study was approved by the Canadian Council for Animal Care and conducted at the Research Center of the University of Montreal Hospital Center (CRCHUM).

Wild-type zebrafish (Danio rerio) were maintained at 28.5°C and kept under a 12/12 h light/dark cycles at the animal facility of the Centre Hospitalier de Montréal research center, Montréal, Canada. They were bred according to standard procedure (Westerfield, 1995) and staged as previously described (Kimmel et al., 1995).

The pCS2-TDP43-Myc and pCS2-TDP43-G348C-Myc plasmids were generously shared by Dr. Guy Rouleau (McGill University). They respectively contain the human wild type TARDBP cDNA and the human TARDBP cDNA bearing the G1176T change encoding for the G348C mutation. Six sequences coding for the myc tag were added to the 3′ end of the sequences. The Tol2Kit was used to generate the two pDest-Tol2CG2-hsp70-TDP43-Myc and pDest-Tol2CG2-hsp70-TDP43-Myc vectors. 25 ng/μl of these vectors were injected in one-cell stage zebrafish embryos with 25 ng/μl of the Tol2 transposase mRNA with 0.05% Fast Green (Sigma). The injections were done using a Picospritzer III pressure ejector (General Valve, Fairfield, NJ, USA). Approximately 300 embryos were injected for each line and < 40% expressed eGFP in the heart. These embryos were selected and raised to adulthood as the F0 founders. For the TDP-43^WT line, 25 F0 fish reached sexual maturity and 11 of them transmitted the transgene to their progeny when outcrossed with wild type AB zebrafish, as identified by eGFP expression in the heart. The percentage of transgene transmission for these F0 founders varied from 0.69 to 33.54%. We selected the two F0 fish with the highest transgene transmission rate (28.44 and 33.54%) to raise the F1 generations. The adult F1 TDP-43^WT fish were then outcrossed to identify the rate of transgene transmission and estimate the copy number of the transgene. The transmission rate of one of the TDP-43^WT lines followed normal mendelian ratios for a single insertion, with 50% of the progeny inheriting the transgene and we decided to only work with this line. 32 TDP-43^G348C F0 fish reached adulthood and only 7 of them transmitted the transgene to their progeny, at a rate of 0.78–29.42%. Due to the low percentages, only the two F0 fish transmitting with the highest frequency per line (17.95 and 29.42%) were selected to raise their progeny until adulthood, to create the two independent lines A-TDP-43^G348C and B-TDP-43^G348C. The adult F1 fish were then outcrossed to identify the rate of transgene transmission and both TDP-43^G348C lines displayed mendelian ratios of transgene transmission, with around 50% of transgene transmission.

For the heat shock procedure, 20 hpf embryos were transferred to glass tubes with system water. The tubes were then placed in a pre-heated incubator at 38.5°C for 60 min at a 60 RPM agitation. Embryos were then put back at 28.5°C and dechorionated at 25 hpf.

Western blots were performed as previously described (Gan-Or et al., 2016). Briefly, about 30 embryos for each condition were lysed in ice-cold RIPA buffer (150 mM NaCl, 50 mM Tris pH 7.5, 1% Triton X-100, 0.1% SDS, 1% Na deoxycholate, 0.1% protease inhibitor), maintained on ice and homogenized with a hand-held pestle. The lysates were centrifuged for 10 min at 10,000 rpm at 4°C and the supernatants were collected. Protein concentration was established using Bio-Rad DC Protein Assay and 60 μg of proteins were loaded in 2X Lämmli buffer after boiling the samples for 5 min at 95°C. TGX Stain-Free FastCast acrylamide Kit (Bio-Rad) was used to make 10% Tris-glycine. After electrophoresis, the gels were activated by exposure to UV for 1 min, and transferred to either regular or low-fluorescence PVDF membrane (Bio-Rad). After transfer, total proteins on membranes were detected by UV. Membranes were then blocked for 1 h using 5% fat free milk in PBST. The primary antibody (anti-myc, 1:5,000) was incubated overnight at 4°C in 5% fat free milk in PBST. After washing 3 × 5 min in PBST, the membrane was incubated with an anti-mouse IgG, HRP conjugated (1:5,000, Jackson Immuno) in 5% fat free milk in PBST for 1 h at room temperature, and then washed 3 × 5 min. Membranes were exposed to Clarity enhanced chemiluminescence (ECL, Bio-Rad) for 5 min at room temperature and visualized using a ChemiDoc MP (Bio-Rad). Detection and quantification of band intensities was conducted using Image Lab 5.2 software (Bio-Rad).

Whole-mount immunohistochemistry was done like previously described (Dzieciolowska et al., 2017). Briefly, 2 dpf embryos were fixed in 4% paraformaldehyde overnight at 4°C. Following fixation, the embryos were rinsed several times over the course of an hour with phosphate buffered saline and incubated in 1 mg/ml collagenase in PBS for 40 min. The collagenase was washed off over 1 h with PBS and the embryos were then incubated with Triton X-100 (PBST) for 30 min. Following this step, the embryos were incubated in fresh block solution prepared from PBS containing goat serum, bovine serum albumin, dimethyl sulfide (DMSO) and Triton X-100, for 1 h at room temperature and then treated with a solution containing a primary antibody against pre-synaptic synaptotagmin 1 (znp-1) (1:100; Molecular Probes) overnight at 4°C. Samples were then washed in PBST and incubated in block solution containing a secondary antibody (Alexa Fluor 488, 1:1000; Invitrogen) for 6 h at 4°C. Before imaging, larvae were transferred to a solution containing 70% glycerol and mounted the following day on a slide. The embryos were visualized using a Quorum Technologies spinning disk confocal microscope with a CSU10B (Yokogawa) spinning head mounted on an Olympus BX61W1 fluorescence microscope and connected to a Hamamatsu ORCA-ER camera. Images were acquired using Volocity software (Improvision) and analyzed using ImageJ (NIH).

Locomotor activity was assessed by lightly touching the tail of the zebrafish embryos with a light forceps. The embryos were placed at the center of a 150 mm diameter petri dish filled with system water that was kept at 28.5°C by a heating plate. Swimming was recorded from above at 30 Hz (Grasshopper 2 camera, Point Gray Research) and swim duration, swim distance and maximum swim velocity were quantified off line using the Manual Tracking plug-in for ImageJ (NIH).

Total RNA from around 100 RNAlater-fixed embryos (Ambion) was extracted using RNAsolv reagent (Omega Biotek) following the manufacturer's standard protocol. RNA extraction was made using between 5.19 and 9.75 × 105 cells by RNAsolv reagent manufacturer's protocol (Omega Biotek). Absence of contamination with chemicals was assessed by nanodrop using 260/280 and 260/230 ratios. Quantification of total RNA was made by nanodrop and 0.1–1.44 g of total RNA was used for sequencing. Quality of total RNA was assessed with the BioAnalyzer Nano (Agilent).

Library preparation was done with the Truseq RNA (Illumina). 18 PCR cycles was required to amplify cDNA libraries. Libraries were quantified by nanodrop and BioAnalyzer. All librairies were diluted to 10 nM and normalized with the Miseq SR50 v2. Libraries were pooled to equimolar concentration and multiplexed by six samples per lane. Sequencing was performed with the Illumina Hiseq2000 using the SBS Reagent Kit v3 (100 cycles, paired-end) with 1.6 nM of the pooled library. Cluster density was targeted at around 800 k clusters/mm². Around 70 million reads were generated for each sample. Library preparation and sequencing was done at the Institute for Research in Immunology and Cancer's Platform (University of Montreal). About 95% of high quality reads were mapped onto GRCz10 version of the zebrafish genome (gene annotation from Ensembl version 87) using STAR version 2.5.1b.

Differential gene expression analysis was assessed by DeSeq package using R software. Differential gene expression was filtered on a False Discovery Rate (or adjusted p-value) < 0.05.

Quantitative RT-PCR (qRT-PCR) primers for each gene were designed using the Universal Probe Library tool from Roche. Total RNA was extracted from FACS-purified NSCs using RNAsolv reagent (Omega Biotek) and chloroform followed by isopropanol precipitation. Reverse transcription was performed from 1 μg of total RNA using the superscript VILO reverse transcription mix (Invitrogen). Quantitative PCR was performed on 2 μL of 1:10-diluted cDNA using SYBR Green I master (Roche) on a LightCycler 80 thermocycler. polr2d gene (ENSDART00000108718) was used as a reference gene for ddCt quantification.

In order to study the effects of mutant TDP-43, we have generated transgenic zebrafish lines expressing stably either human wild-type TDP-43 (TDP-43^WT) or bearing the G348C mutations found in ALS patients from a French-Canadian population (TDP-43^G348C) (Kabashi et al., 2008). The transgenic lines were generated using the human TDP-43 cDNA to which 6 myc tags were fused to the C-terminus of the protein. Since previous work showed that the presence of mutant TDP-43 in the early stages of development of the zebrafish lead to adverse effects (Kabashi et al., 2010, 2011; Laird et al., 2010; Vaccaro et al., 2012, 2013; Armstrong and Drapeau, 2013; Patten et al., 2017), we decided to use an inducible promoter, the zebrafish heat shock protein 70 like promoter (hsp70l), to temporally control the expression of the transgene. The transgene constructs also contained eGFP driven by the cardiac myosin light chain 2 (cmlc2) promoter used as a reporter gene in order to easily identify transgenic embryos with green hearts (Figure 1A).

The constructs were injected in 1–4 cell stage embryos (F0) and the fish were raised until adulthood. The adult F0 fish were outcrossed with wild-type fish and their progeny was screened for the presence of fluorescent hearts, to select the ones having integrated the transgenes in their germline. We selected the two F0 fish with the highest transmission rate for both lines to raise the F1. These F1 fish were outcrossed with wild-type to estimate the transgene copy number by the transgene transmission rate. While both TDP-43^G348C lines had an ~50% transmission rate of the transgene, only one of the TDP-43^WT line transmitted the transgene following mendelian ratios and only this line was kept for further experiments. We selected the TDP-43G348C line (B) that had the closest level of expression to the TDP-43WT line (Supplementary Figure S1). The F2 offspring were raised to adulthood and used to establish the lines for the following experiments.

In order to express the transgene, a 60 min heat shock at 38.5°C (10°C above the temperature at which they are raised) is done when the embryos are 20 h post-fertilization (hpf). Only normal looking embryos are selected at 48 hpf for experiments (Figure 1B). To establish the expression of the transgenes by Western blot, we used an anti-myc antibody. No transgenic protein was detectable before the heat shock, thus validating the integrity of the hsp70l promoter. However, upon heat shock the TDP-43^WT and TDP-43^G348C transgenes were detectable at around 60 kDa, which corresponds to the molecular weight of the TDP-43 protein with the linker and the 6 myc tags (Figure 1C). Interestingly, we observed the TDP43^WT expression level was systematically lower than that of TDP-43^G348C. Moreover, several bands were also observed in the TDP-43^G348C line, indicating the presence of C-terminus TDP-43 fragments, consistent with what is seen in human patients and in mouse studies (Stallings et al., 2010; Swarup et al., 2011; Xu et al., 2011; Gao et al., 2018) (Figure 1C).

We then verified if the stable expression of TDP-43^G348C upon heat shock was associated with locomotor defects, as it was observed in previous studies by transient expression. We elicited the touch-evoked response by lightly touching the tail of 48 hpf embryos with a blunt forceps. The swim distance, swim duration and the maximum swim velocity were analyzed using the swim traces obtained by tracking the embryos following the touch (Figure 2A). While heat shock of wild type embryos or expressing TDP-43^WT had no influence on locomotion compared to wild type, the expression of TDP-43^G348C led to a strong reduction of the swim distance, the swim duration and the maximum swim velocity compared to controls (Figure 2B), indicating that the motor neuron axonal defects had a strong impact on the physiology as seen by the severe negative influence on the locomotion of the TDP-43^G348C embryos, as was previously reported (Kabashi et al., 2010, 2011; Laird et al., 2010; Vaccaro et al., 2012, 2013; Armstrong and Drapeau, 2013; Patten et al., 2017).

As locomotor defects were observed, we investigated whether they reflected axonal motor neuron defects by immunohistochemistry using an antibody against the pre-synaptic marker znp-1. When TDP-43^G348C was expressed, the majority of motor neuron axons exhibited several defects, including overbranching of the main axonal branch, absence of secondary branching or abnormal innervation of nearby myotome (Figure 3A). The axonal defects were classified as mainly moderate (Figure 3C) using a classification scale based on the visual interpretation of the axonal appearance, adapted from McWhorter et al (Figure 3B) (McWhorter et al., 2008). The overexpression of the TDP-43^WT transgene, or the heat shock of wild-type embryos did not induce specific defects, indicating that it was the specifically the presence of mutant TDP-43 rather than the heat shock or the presence of a foreign gene that was responsible for the defects obtained with TDP-43^G348C. This is consistent with previous zebrafish models overexpressing mutant TDP-43 (Kabashi et al., 2010, 2011; Laird et al., 2010; Vaccaro et al., 2012, 2013; Armstrong and Drapeau, 2013; Patten et al., 2017). Altogether, these data validate the relevance of our stable expressing transgenic zebrafish lines since they display a clinically-relevant pathogenicity associated with the G348C mutation.

Given that TDP-43 is an RNA-binding protein involved in RNA metabolism (Gao et al., 2018), we next sought to investigate whether transcriptomic modifications were caused by the expression of TDP-43^G348C. While TDP-43 is involved in mRNA splicing, it is also involved in mRNA stability and the depletion of Tdp-43 causes around 600 mRNAs to have altered expression levels in mice (Polymenidou et al., 2011). Thus, we used RNA-sequencing to look at differentially regulated genes due to TDP-43^G348C expression. Since the human TDP-43^WT line following heat shock presented no obvious phenotype compared to wild-type zebrafish with heat shock, and since the expression level of human TDP-43^WT was lowered than human TDP-43^G348C with an equal heat shock, we hypothesized that using human TDP-43^WT as a control may give us differentially expressed genes based on the expression level of the transgenes rather than solely based on the presence of the mutation. We thus decided to use the wild type zebrafish heat shock condition as a control in order to compensate for the effect of the heat shock.

More than 70 million reads were sequenced and analyzed for either TDP-43^G348C heat shock or wild type heat shock and more than 95% were mapped to the GRCz10 version of the zebrafish genome (gene annotation from Ensembl version 87). We used DESeq to analyze the differentially expressed genes amongst the 32,267 genes analyzed between the two conditions. Filtering on a p-value < 0.05 we found 159 differentially expressed genes, with 67 up-regulated and 92 down-regulated. In order to have a more stringent cut-off, we used a |log2FC| greater than one, corresponding to a minimal 2-fold up-regulation or down-regulation. This cut-off decreased the number of differentially expressed genes to 148 (65 up-regulated and 83 down-regulated). An even more stringent cut-off of |log2FC| greater than two, corresponding to a minimal 4-fold up- or down- regulation, led to only 59 genes that were differentially regulated, with 27 up-regulated and 32 down-regulated (Figure 4A). A volcano plot representing these data is shown in Figure 4B with data in, blue, orange or black for cut-offs of |log2FC| greater than two or one or for p < 0.05, respectively. The list of significantly differentially expressed genes can be found in the Supplementary Table S1.

These results show that the expression of a G348C mutant TDP-43 is associated with broad changes in the transcriptome. While no common pathways could be found between the 159 genes with p < 0.05, some of these changes can be associated with potential pathophysiological mechanisms of ALS and TDP-43 mutations. Thus, the human orthologs of some of the differentially expressed genes have been linked to several neuromuscular disorders, including ALS and muscular dystrophy, such as CLCN2, ELOVL4, ELOVL5, PEA15, LBH, PAMR1, and MSTO1 (references in Table 1). Other genes can be linked to pathways that are perturbed in ALS, such as calcium signaling (icn2, smdt1a, pvalb4, pvalb9 and adcy1b), or mitochondria and oxidative stress (msto1, smdt1a, and mt2). Also of interest, some of the genes identified are unknown, such as ponzr3, ponzr4, or ftr86, which may indicate new pathogenic mechanisms of ALS (references in Table 1).

In order to validate these results, we selected 12 of the differentially expressed genes to measure their expression levels by qRT-PCR in all of our conditions, including TDP-43^WT heat shock (Figures 5A–L). All mRNA expression levels positively correlated with the RNA-sequencing data, even though significance was not always reached (Figures 5D–G,J,K), probably due to variability in the level of mRNA expression (Figures 5D,F,I,J).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.