Effect of Ataluren on dystrophin mutations

Abstract Duchenne muscular dystrophy is a severe muscle wasting disease caused by mutations in the dystrophin gene (dmd). Ataluren has been approved by the European Medicines Agency for treatment of Duchenne muscular dystrophy. Ataluren has been reported to promote ribosomal read‐through of premature stop codons, leading to restoration of full‐length dystrophin protein. However, the mechanism of Ataluren action has not been fully described. To evaluate the efficacy of Ataluren on all three premature stop codons featuring different termination strengths (UAA > UAG > UGA), novel dystrophin‐deficient zebrafish were generated. Pathological assessment of the muscle by birefringence quantification, a tool to directly measure muscle integrity, did not reveal a significant effect of Ataluren on any of the analysed dystrophin‐deficient mutants at 3 days after fertilization. Functional analysis of the musculature at 6 days after fertilization by direct measurement of the generated force revealed a significant improvement by Ataluren only for the UAA‐carrying mutant dmdta222a. Interestingly however, all other analysed dystrophin‐deficient mutants were not affected by Ataluren, including the dmdpc3 and dmdpc2 mutants that harbour weaker premature stop codons UAG and UGA, respectively. These in vivo results contradict reported in vitro data on Ataluren efficacy, suggesting that Ataluren might not promote read‐through of premature stop codons. In addition, Ataluren had no effect on dystrophin transcript levels, but mild adverse effects on wild‐type larvae were identified. Further assessment of N‐terminally truncated dystrophin opened the possibility of Ataluren promoting alternative translation codons within dystrophin, thereby potentially shifting the patient cohort applicable for Ataluren.


| INTRODUC TI ON
MD treatment is the corticosteroid prednisone (prednisolone), which only mitigates the rate of muscle degeneration. 1 In addition, the exon-skipping drug Eteplirsen has been approved by the Food and Drug Administration of the USA (FDA) but not the European Medicines Agency (EMA). 2 In 2007, Ataluren (Translarna, PTC124) was identified in a high-throughput small molecule screen and suggested to promote suppression of nonsense mutations. 3 In cell culture experiments, the efficiency of Ataluren was reported to inversely correlate with termination efficiencies in vitro; being most efficient with UGA, the most permissive PTC, followed by UAG and then UAA. 3 In addition, oral administration of Ataluren ameliorated the dystrophic condition of the mdx mouse, which harbours the nonsense mutation UAA in Dmd that is least susceptible to Ataluren. 3 Importantly, these findings were questioned by later reports that demonstrated that Ataluren stabilized the firefly luciferase reporter used to identify Ataluren. 4,5 PTCs within dystrophin have been identified as the genetic cause of Duchenne MD in approximately 10%-15% of patients. 6,7 In smallscale as well as in randomized, double-blind, placebo-controlled studies, Ataluren was found to have beneficial effects on disease progression in patients carrying nonsense mutations within dystrophin. [8][9][10][11] However, Ataluren treatment showed a significant clinical benefit only in the subgroup of patients with a baseline 6-minute walk distance of 300 m or < 400 m, no significant effect was recorded for other subgroups. 10 In addition, in studies of Duchenne muscular dystrophy patients, the nonsense mutation type was found to be not associated with beneficial effects of Ataluren. 11,12 These studies also revealed that Ataluren was generally well tolerated and most adverse clinical events were mild to moderate in severity, 10,12 including a possible reduction in body mass index. 8 Therefore, the discussion about Ataluren efficacy on Duchenne MD patients with nonsense mutations as well as the mechanism of Ataluren function has not been resolved. 13 As a result, Ataluren has been recently approved by the EMA but not the FDA.
Zebrafish are a valuable model to study Duchenne MD due to their close replication of the human condition, 14 as well as their fecundity and genetic susceptibility that enables rigorous phenotypic evaluation. In addition, robust assays have been established that allow quantification of muscle parameters within zebrafish larvae.
The birefringence assay employs the muscle birefringence, a light effect provoked by the diffraction of polarized light through the pseudo-crystalline array of the muscle myofibril, causing muscle fibres to appear bright in an otherwise dark environment. 15 Therefore, dystrophic muscle of dystrophin-deficient mutants that feature stochastic myofibre breakdown display a patchy birefringence pattern and their phenotypic severity can be quantified at 3 days after fertilization (dpf). A further assay directly measures the force generated by whole mounted zebrafish larvae, 16 thereby enabling quantification of muscle parameters in zebrafish models of muscle diseases. 17,18 Importantly, Ataluren administration to the dystrophin-deficient dmd ta222a zebrafish significantly ameliorated the force generation of homozygous larvae and restored low levels of dystrophin protein. 19 In this study, additional zebrafish dystrophin-deficient lines were generated to enable evaluation of Ataluren efficiency in relation to different PTCs. In an extensive series of Ataluren treatments, only dmd ta222a homozygotes were ameliorated by Ataluren. Interestingly, the TAA nonsense mutation within dmd ta222a has been reported to be least effective for Ataluren treatment. 3 The other tested mutants carrying TAG and TGA nonsense mutations or a frame-shifting deletion remained unaffected, indicating that Ataluren might not function by ribosomal PTC read-through. In addition, Ataluren had mild adverse effects on wild-type (WT) larvae, which is consistent with human trials. 8,10 Digital droplet PCR revealed that Ataluren had no effect on dystrophin transcript in dmd ta222a mutants and WT siblings, indicating that the ameliorative effect of Ataluren was not conveyed via dystrophin transcript. In contrast to the other mutants that carry downstream mutations (dmd pc2 in exon 32, dmd pc3 in exon 34), the mutation of dmd ta222a locates to exon 4. Transgenic assessment of dystrophin protein showed that N-terminally truncated dystrophin protein lacking exons 1-7 was able to partially rescue the dystrophic phenotype of dystrophin-deficient zebrafish. This opens the possibility that, instead of supressing PTC, Ataluren might promote translation of dystrophin from alternative start codons. However, further insights into Ataluren function are required to reveal its mechanism of action and to identify the patient cohort responsive to Ataluren treatment.

| Generation and genotyping of zebrafish mutant lines
A total of 48 males were mutagenized with N-ethyl-N-nitrosourea (ENU) as described before 20 and approved by the Monash Animal Service (MAS/2009/05). Surviving fish were outcrossed, and obtained F1 fish were crossed to heterozygous dmd ta222a . Resulting offspring was analysed for complementation via the birefringence assay at 3 days after fertilization (dpf). This non-complementation screen identified the dmd pc3 dystrophin allele, which was subsequently genotyped by PCR with the primers cDMD_DdeI_F (5′-ccgctcatcggtagagggtatgccaagtgctt) and gDMD_DdeI_R (5′-gagcactacaatcagtgaatgataacaa) followed by restriction digestion with DdeI (NEB).

| Ataluren treatment
Ataluren (Selleck Chemicals) was dissolved as a stock solution of 10 mmol/L in Dimethyl sulfoxide (DMSO) and in experiments directly added to the fish water at a final concentration 0.5 µmol/L. The Ataluren concentration of 0.5 µmol/L was established as optimal effective concentration within zebrafish previously. 19 The control group was treated with 0.005% DMSO as a negative vehicle control, thereby matching the DMSO concentration used with Ataluren-treated fish.
Zebrafish embryos were dechorionated at 24 hpf before treatment.
Fish water solutions were exchanged on a daily basis until analysis.

| Quantification of birefringence
At 72 hpf, individual zebrafish larvae were automatically imaged in an unbiased way using the Abrio LS2.2 microscope as previously described. 15 To maximize uniformity of larvae stages, all larvae of an analysed clutch were kept at 25°C during the imaging process in order to slow down larvae development without adverse effects. 22 In addition, all imaging was performed within less than 1 hour.
Subsequently, the first 20 somites of imaged larvae were selected using the software ImageJ and the mean of all grey values of the pixels was measured. To enable comparison of the birefringence from different larvae, obtained grey values were rescaled to control siblings set to 100%. To rescale values of siblings, measured grey values (A 1 to A n ) of each larva were multiplied by 100 and divided by the average of all measured grey values of siblings using To normalize values of mutants, measured grey values (B 1 to B n ) of each larva were multiplied by 100 and divided by the average of the measured grey values of the siblings using clutches, a minimum of five siblings and five mutants were analysed for their muscle birefringence, each treated with either 0.005% DMSO (negative vehicle control) or 0.5 µmol/L Ataluren as indicated. Finally, all analysed animals were genotyped by PCR as described above.

| Statistical analysis
Statistical significance was calculated using the software Prism (GraphPad Software). Between two groups, significance was determined by Student's t test and for multiple groups one-way ANOVA with post hoc Tukey's test was used. Presented data are mean ± SEM, calculated utilizing error propagation.

| Force measurement
6-dpf-old larval were individually mounted at slack length between a force transducer and a puller with aluminium clips as described earlier. 23 In short, whole larvae preparations were kept at 22°C in physiological buffered solution and stimulated through electrical pulses of 0.5 ms duration (supramaximal voltage) to give single twitches.
Isometric force analysis was performed at various larva lengths to identify the optimal muscle length for maximal active force quantification. Single twitches were separated in 2-min intervals and performed at stepwise-increased length. At each length, active contraction was recorded to identify the maximal active force at optimal length. Subsequent to analyses, all animals were genotyped by PCR as described above.
Animal experiments were approved by IBC/22219.
Subsequently, 70 µL droplet generation oil (Bio-Rad) was added to the ddPCR reaction and the Bio-Rad QX200 droplet generator (Bio-Rad) was utilized to generate up to 20 000 droplets from each sample. After PCR amplification, the QX200 Droplet Reader (Bio-Rad) was used for automatic readout and results were analysed with the QuantaSoft software (Bio-Rad).

| Generation of novel dystrophin-deficient zebrafish mutants
The dystrophin-deficient zebrafish mutant line dmd ta222a harbours a nonsense mutation in exon 7 that encodes the PTC UAA 28 and dmd pc2 carry UGA in exon 32. 20 To assess the reported inverse correlation of Ataluren efficacy with termination efficiencies of the three different PTCs in an animal model of Duchenne MD, novel zebrafish mutants were generated. In a genetic non-complementation screen, male zebrafish were outcrossed after ENU treatment and resulting F1 founders were crossed to heterozygous dmd ta222a/+ fish. The resulting offspring was assessed for complementation of the dmd ta222a allele by birefringence analysis at 3 dpf. This genetic non-complementation approach resulted in the identification of a novel dystrophin mutant, dmd pc3 , that harboured an in-frame PTC within exon 34 ( Figure 1A) leading to loss of dystrophin protein ( Figure 1B). Myofibre detachment, typical for dystrophic muscle, was demonstrated within dmd pc3 in the transgenic background of Tg(acta1:mCherryCaaX) and Tg(acta1:lifeact-GFP), in which mCherry-CaaX highlights the sarcolemma together with t-tubules and Lifeact-GFP directly marks actin thin filaments ( Figure 1C). 29 To generate a dystrophin mutant with a frameshift allele, which should not be affected by PTC suppression, the CRISPR/Cas9 technology was employed. Two single guide RNAs targeting exon 53 and 53/54 intron were co-injected with Cas9 into WT eggs. After germline transmission, the novel dmd −69bp allele was identified in which the 3′ splice site of exon 53 was removed ( Figure 1D). The altered splicing of the dystrophin transcript in dmd −69bp homozygotes led to a frameshift and multiple subsequent PTCs within the dystrophin coding sequence and loss of dystrophin protein ( Figure 1D,E). Myofibre retraction within dmd −69bp homozygotes was confirmed in the double transgenic background of Tg(acta1:mCherryCaaX) and Tg(acta1:lifeact-GFP) ( Figure 1F).
In summary, two novel dystrophin-deficient mutants dmd pc3 and dmd −69bp were generated: dmd pc3 with the PTC UAG and F I G U R E 1 Novel dystrophin-deficient mutants. A, A > T substitution within exon 34 of dmd pc3 results in a nonsense mutation. B, Wholemount immunohistochemistry with antibodies against dystrophin revealed loss of dystrophin protein in 3-dpf-old dmd pc3 homozygotes. C, At 3 dpf, labelling of the sarcolemma with mCherryCaaX and the myofibril with Lifeact-GFP confirmed myofibre detachment within dmd pc3 homozygotes. D, 5 bp from exon 53 and 64 bp from the downstream intron was removed from the dystrophin gene in dmd −69bp mutants. Altered splicing in dmd −69bp led to integration of 10 bp from the intron downstream of exon 53 into the dystrophin transcript, resulting in a frameshift and multiple subsequent PTCs. E, Dystrophin protein is lost in dmd −69bp homozygotes, as indicated by antibodies against dystrophin at 3 dpf. F, Retracting myofibres within 3-dpf-old dmd −69bp were revealed in the double transgenic background of Tg(acta1:mCherryCaaX) and Tg(acta1:lifeact-GFP)

Genomic
Transcript with a deletion not susceptible to PTC readthrough.
Both mutants phenotypically match previously obtained dystrophin mutants, while their siblings remain phenotypically unremarkable.

| Ataluren treatment over 2 days does not significantly affect the muscle integrity of dystrophindeficient zebrafish mutants
To

| Ataluren treatment over 5 days significantly ameliorates only dmd ta222a homozygotes that feature a nonsense mutation in exon 4 of dystrophin
Due to the growing muscle thickness, the birefringence assay has to be employed at early larvae stages. In order to assess the effect of Ataluren treatment over 5 days and analyse muscle function in addition to the muscle pathology at 3 dpf, the maximal force generated by 6-dpf-old larvae was measured. At 24 hpf, zebrafish embryos were dechorionated and exposed to either 0.5 µmol/L Ataluren or DMSO vehicle control. Solutions were renewed on a daily basis, and larvae were separated according to their phenotype at 3 dpf. At 6 dpf, 4 siblings and 4 homozygotes of each treatment group were randomly selected and subjected to force measurement ( Figure 3). Subsequently, the genotype of all larvae was verified by PCR-based genotyping. Maximal force measurements using a force transducer revealed that Ataluren treatment significantly improved the maximal force generation of homozygotes dmd ta222a larvae compared with non-treated homozygotes (Figure 3), which was consistent with reported results demonstrating that Ataluren treatment ameliorated force generation and restored dystrophin expression in dmd ta222a homozygotes. 19

| Ataluren treatment has a mild, but significantly adverse effect on zebrafish
Interestingly, assessment of Ataluren treatment of four different dystrophin-deficient zebrafish lines revealed that administration of 0.5 µmol/L Ataluren had a mild adverse effect on all tested sibling groups, although this tendency was not significant. However, by pooling the results obtained from all wild-type (WT) larvae from the four analysed dystrophin-deficient lines, the effect of Ataluren becomes significant (Figure 4). The birefringence of 3-dpf-old WT larvae that were treated with 0.5 µmol/L Ataluren over 2 days was significantly reduced in comparison to DMSO-control WT larvae, demonstrating that Ataluren leads to a reduction in the amount of myofibril ( Figure 4A). Similarly, the maximal force generated by 6-dpf-old WT larvae was significantly reduced after 5 days of 0.5 µmol/L Ataluren treatment compared with DMO-treated WT controls ( Figure 4B).
This result revealed that Ataluren had a mild, but significant adverse effect on the musculature of healthy wild-type zebrafish larvae.

| Dystrophin Transcript levels are not affected by Ataluren treatment
To assess whether the significant amelioration of the muscle force generated by dmd ta222a mutants after Ataluren treatment was based on elevated levels of dystrophin transcript, droplet digital PCR (ddPCR) was employed. Droplet digital PCR (ddPCR) has emerged as a reliable analytical technology for sequence-specific detection and precise quantification of nucleic acids, facilitating reproducible measurement of small percentage differences even of rare variants. 30 Similar to the force measurement assay, 24-hpf-old zebrafish embryos were dechorionated and exposed to either 0.5 µmol/L Ataluren or DMSO vehicle control. At 6 dpf, larvae were genotyped and two larvae per genotype (WT siblings or dmd ta222a homozygotes) and treatment group (DMSO control or Ataluren) were pooled into biological samples. Subsequently, the quantity of dmd transcript from three biological replicates of each of the four treatment groups was measured. In a one-step reverse-transcription ddPCR, dystrophin transcript levels were quantified relative to transcript levels of the polr2d reference gene. Compared with DMSO control-treated wild-type siblings, a highly significant reduction in the relative amount of dmd transcript was detected in control-treated dmd ta222a homozygotes ( Figure 5), likely caused by non-sense mediated decay of mutant dystrophin transcript. 31  with control-treated homozygotes remained unchanged. Similarly, dmd transcript within wild-type siblings was not affected by Ataluren ( Figure 5).
Thereby, a significant effect of Ataluren on the level of dystrophin transcript was not detected, indicating that the effect of Ataluren on dmd ta222a might not be conveyed via dystrophin transcript.

| N-terminally truncated dystrophin significantly ameliorates the phenotype provoked by lack of endogenous full-length dystrophin
Enhancement of the maximal force generated by dmd ta222a homozygotes and restoration of dystrophin protein by Ataluren has been demonstrated. 19 Figure 6A).
Importantly, GFP fluorescence within Tg(cry:GFP,acta1:dmdGFP) was not only detected in the lens but also at the myotendinous junctions, indicating that Dmd-GFP fusion protein replicated the localization of endogenous dystrophin ( Figure 6B).
To analyse the rescue of the dystrophic phenotype of dystrophin mutants lacking endogenous dystrophin, dmd pc2 was crossed to into the Tg(cry:GFP,acta1:dmdGFP) transgenic background and their birefringence was assessed at 72 hpf ( Figure 6C). In contrast to non-transgenic dmd pc2 homozygotes that featured a highly significant reduction in birefringence compared with their non-transgenic siblings, the birefringence of dmd pc2 homozygotes in the Tg(cry:GFP,acta1:dmdGFP) background was comparable to non-transgenic siblings. The rescue of dmd pc2 homozygotes indicates that the transgenic full-length dystrophin, fused to GFP and expressed from a transgene under the acta1 promoter, was fully functional ( Figure 6D).
In transgenic zebrafish, dystrophin lacking exons 1-7 was able to ameliorate the dystrophic pathology. One could therefore speculate that Ataluren could enhance translation of dystrophin transcript from alternative start codons.

| D ISCUSS I ON
The efficacy of Ataluren remains disputed, and its mechanism of action has not been resolved utilizing animal models and cultured cells. Additionally, Ataluren assessment in patients suffering from C, Polarized light visualized the muscle of 3-dpf-old larvae. Whereas the birefringence of non-transgenic siblings appeared uniform, the birefringence of dmd pc2 homozygotes appeared patchy. In contrast, in the transgenic background of Tg(cry:GFP,acta1:dmdGFP) the birefringence of dmd pc2 homozygotes and siblings was comparable. D, Quantification of the birefringence followed by rescaling to nontransgenic siblings revealed that the birefringence of dmd pc2 homozygotes transgenic for Tg(cry:GFP,acta1:dmdGFP) was not significantly changed compared with non-transgenic dmd pc2 siblings. Crosses represent averaged grey values of at least five 72-hpf-old larvae. Black bars represent mean ± SEM and n.s. non-significant. Significance was determined by one-way ANOVA with Tukey's multiple comparisons post hoc test (***P < 0.001, n = 6). E, Within Tg(cry:GFP,acta1:dmd ∆ex1-7 GFP) larvae, N-terminally truncated dystrophin fused to GFP localized to the vertical myosepta as indicated by the GFP fluorescence. F, After rescaling to non-transgenic siblings, the birefringence of dmd pc2 homozygotes transgenic for Tg(cry:GFP,acta1:dmd ∆ex1-7 GFP) was significantly ameliorated compared with non-transgenic dmd pc2 homozygotes, but significantly reduced compared with non-transgenic dmd pc2 siblings. Crosses represent averaged grey values of at least five 72-hpf-old larvae. Black bars are mean ± SEM Significance was determined by one-way ANOVA with Tukey's multiple comparisons post hoc test (***P < 0.001, n = 6). G, At 6 dpf, the maximal force generated by dmd pc2 homozygotes positive for Tg(cry:GFP,acta1:dmd ∆ex1-7 GFP) was significantly stronger compared with non-transgenic dmd pc2 homozygotes, but significantly reduced compared with non-transgenic dmd pc2 siblings. Crosses represent individual larvae and black bars mean ± SEM. Significance was determined by one-way ANOVA with Tukey's multiple comparisons post hoc test (**P < 0.01, n = 4) and UGA that, based on reported data, 3 are predicted to be more susceptible to Ataluren. These findings are consistent with smallscale clinical studies of Ataluren, in which only a subgroup of nonsense mutation Duchenne muscular dystrophy patients benefitted and no correlation with PTC types was found. 10,12 Although PTC suppression by Ataluren has been confirmed in vitro, 34 human trials and our zebrafish study, these studies combined indicate that additional factors might contribute to the beneficial effect of Ataluren. A potential contributing factor might be the complex influence of nucleotides downstream of the PTC that have been reported to alter the termination efficiency of PTCs. 35 Pooling of results obtained from wild type revealed that Ataluren had mild detrimental effects on the muscle of zebrafish larvae, which showed a reduction of the amount of myofibril after 3-day treatment and weakening of the muscle after 5-day treatment. This finding is in agreement with results from clinical trials involving Duchenne MD patients, reporting mild to moderate adverse clinical events. 8,10,12 Accordingly, emerging evidence has been brought forward that lower doses of read-through supressing aminoglycosides increased misincorporation of amino acids during translation 36 and that aminoglycosides induced damage to the kidney and the inner ear. 37,38 Therefore, although Ataluren was generally well tolerated, 8,10 caution is required when Ataluren is provided to patients.
To test the possibility of Ataluren enforcing translation from alternative ATG codons, the functionality of N-terminally truncated dystrophin was assessed in transgenic animals. Partial functionality of dystrophin lacking exons 1-7 was indicated by the partial rescue of the dystrophic muscle of dmd pc2 homozygotes. These findings are consistent with patient reports revealing a nonsense mutation within exon 1 resulting in the mild Becker MD symptoms due to the alternative translation initiation at two AUG codons located in exon 6, 39 which are employed by internal ribosomal entry site within exon 5. 40 Similarly, for the frame-shifting deletion of exons 3-7 that result in Becker MD, it has been proposed that an alternative ATG in exon 8 could be used for translation initiation. 41 These findings open the possibility of another route of mechanism for the action of Ataluren, in which alternative translation starts are employed to generate shorter but largely functional dystrophin protein. However, whether truncated dystrophin protein was generated in dmd ta222a homozygotes after Ataluren treatment has not been assessed directly.
Duchenne MD trials and our zebrafish study indicate that the mechanism of action of Ataluren has not fully been uncovered. Our additional results from transgenic zebrafish opened the possibility that Ataluren might contribute by promoting dystrophin translation from alternative translation start codons. However, further insights are required to fully establish the mechanism of action of Ataluren in order to identify the cohort of patients responsive to Ataluren.

ACK N OWLED G M ENTS
We are grateful to Monash Micro Imaging of Monash University for technical support. PDC and JB were supported by the National Health and Medical Research Council of Australia (APP1144159 and APP1136567). JB is supported by Muscular Dystrophy Australia.
The Australian Regenerative Medicine Institute is supported by grants from the State Government of Victoria and the Australian Government.

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest.

AUTH O R CO NTR I B UTI O N S
JB conceived the study and performed experiments. ML and SB performed force measurements. MM and JR performed ddPCR. JB and PDC wrote the manuscript, which was approved by all authors.

DATA AVA I L A B I L I T Y S TAT E M E N T
Raw data can be obtained from the corresponding authors on request.