Exon skipping for Duchenne muscular dystrophy: a systematic review and meta-analysis

Background Exon skipping has been considered a promising therapeutic approach for Duchenne muscular dystrophy (DMD). Eteplirsen received conditional approval in the United States in 2016. To date, no systematic reviews or meta-analyses of randomized controlled trials (RCTs) of exon skipping drugs have been published to determine the pooled estimates for the effect of exon skipping in treating DMD. Methods A systematic review and meta-analysis of double-blind RCTs comparing exon-skipping drugs with placebo in DMD was performed. Trials were identified by searching published and unpublished studies from electronically available databases and clinical trial registries through October 2017. The primary outcomes were changes in the 6-min walk test (6MWT) distance, North Star Ambulatory Assessment (NSAA) scores, and adverse events. Random-effects meta-analysis and assessment of risk of bias were performed. This systematic review was registered at PROSPERO (CRD42016037504). Results Five studies involving 322 participants were included, investigating eteplirsen in one and drisapersen in four studies. There were no changes in 6MWT distance (mean difference [MD] − 9.16, 95% confidence interval [CI] − 21.94 to 3.62) or NSAA scores (MD 1.20, 95% CI − 2.35 to 4.75) after 24 weeks of treatment in the exon-skipping group compared with placebo. Subgroup analysis for a 6 mg/kg weekly injection of drisapersen showed significant changes in the 6MWT, favoring drisapersen after 24 weeks (MD − 20.24; 95% CI − 39.59 to − 0.89). However, drisapersen resulted in a significant increase in injection site reactions (risk ratio [RR] 3.67, 95% CI 1.96 to 6.89, p < 0.0001) and renal toxicity (RR 1.81, 95% CI 1.11 to 2.94, p = 0.02). Risk of bias was high in two of the five studies, including the eteplirsen and one drisapersen study. Conclusions Current available data do not show evidence that exon-skipping drugs are effective in DMD. Despite potential effectiveness when used at a specific dose, significant side effects were reported with drisapersen. The small number of RCTs with relatively small numbers of participants indicate the difficulty in conducting sufficiently powered studies of DMD. Prospectively planned meta-analysis and utilization of the real-world data may provide a more precise estimate of the effect of exon skipping in this disease. Electronic supplementary material The online version of this article (10.1186/s13023-018-0834-2) contains supplementary material, which is available to authorized users.


Rationale
3 Describe the rationale for the review in the context of what is already known.

Description of the condition
Duchenne muscular dystrophy (DMD) is caused by the mutation in the DMD located on Xp21 [1], which affects predominantly male individuals. It is the most common hereditary muscular disorder, estimated to affect one in 3500-6000 live male birth [2][3][4]. The protein product of the DMD, dystrophin, links the muscle sarcomeric structure to the extracellular matrix by constructing dystrophin-associated glycoprotein complex (DGC) [5] which confers to the strength of sarcolemma [6]. Individuals with DMD have absent or decreased level of dystrophin, manifesting walking difficulty at around age of five, and they become wheel chair bound before or during their teens [7,8]. The distance of which a subject is able to walk in six minutes assessed by six-minute walk test (6MWT), is interpreted as the measure of physical capacity and walking function [9,10]. In DMD, the longitudinal observational study had shown an increase in the distance of 6MWT until age of 7.5 years, followed by a decline, which became precipitous from 12.5 years to 15.5 years, when all boys were unable to perform the 6MWT [11]. Along with the motor function decline, patients also show respiratory and cardiac dysfunction owing to loss of respiratory muscle strength and cardiac tissue degeneration [12]. Preclinical cardiomyopathy becomes apparent in patients less than age 6 years [13], and respiratory function peaks out before age of 12 [14].
Without supportive care, patients typically die in their late teens or in early 20s [15]. Intervention with mechanical ventilation and cardioprotective medications have improved the survival, and patients may live to their 30s [16].
Currently, there is no curative therapy for DMD and glucocorticoids are the only medication available that slows the decline in muscle strength and function [7]. The clinical efficacy of glucocorticoids is established and is considered to be the standard therapy for patients with DMD, however, there are several concerns over the glucocorticoid therapy. One is that glucocorticoids have serious side effects in long-term usage, including weight gain, cushignoid, behavior change, growth delay, cataracts and osteoporosis [17]; second, the mechanism how glucocorticoids contribute to beneficial effects in patients with DMD is still unknown. Owing to the myriads of side effects, not all patients are eligible for receiving or continuing the glucocorticoid therapy.
Development of innovative therapies for patients with DMD is an urgent task, and several therapeutic approaches are now being investigated, including exon skipping, read through, vector-mediated gene therapy, cell transplantation, anti-inflammatory/fibrotic/oxidant drugs, myostatin pathway inhibitors, nNOS pathway enhancement, and utrophin up-regulation [18,19]. Amongst all these approaches, exon skipping, read through and antioxidants are at the most advanced stage of clinical trials. Two major drugs evaluated for exon skipping are Drisapersen (BioMarin Pharmaceutical Inc.) and eteplirsen (Sarepta Therapeutics), and these have been either completed or now being studied in phase 3 clinical trials (https://clinicaltrials.gov/). Ataluren, the compound for read through has also completed phase 3 study and now has an approval from European Medicines Agency (EMA) for marketing in Europe [20].
Idebenone, an antioxidant has also completed phase 3 trial [21]. Exon skipping and read through aim to treat the disease by restoring dystrophin, whereas idebenone aims to compensate for the lack of dystrophin. Each therapeutic approach bears pros and cons in their nature. Dystrophin is expressed not only in skeletal muscle but also in brain, and have been reported that significant proportion of DMD patients have cognitive and learning disabilities [22]. From the viewpoint that DMD is not only confined to the skeletal muscle but more of a systemic disorder, exon skipping or read through may have more benefit than idebenone.
The idea of exon skipping approach for DMD have emerged in 1990s [23], based on the rationale of converting the translational reading frame from out-of-frame to in-frame to produce shorter but functional dystrophin, instead of degradable dystrophin. Exon skipping can be applicable to versatile mutation types including deletion, duplication, and small mutations, theoretically applicable to up to 83% of total mutation [24], whereas read-through can treat only patients with nonsense mutation which comprises 10% of total mutation [25].
Although it is yet to be clarified whether the efficacy of exon skipping can exceed that of glucocorticoids in DMD patients, exon skipping is one of the most promising approach, with scientific rationale and accumulation of in vitro [26][27][28] and in vivo [29][30][31] study data. However, clinical trials for exon skipping have failed to demonstrate statistically significant improvement in motor function so far [32,33]. Several factors can be thought for the reasons why it is so difficult to replicate pre-clinical research data in patients. One may be the low prevalence of DMD, and that it interferes with designing a large cohort study with sufficient power to detect the efficacy. Hence, a meta-analysis is necessary to assess the efficacy of exon skipping for patients with DMD.
To date, the results from multiple clinical trials have not been reviewed in a standardized manner. Herein, we propose to conduct a systematic review and meta-analysis on exon skipping studies in DMD to assess the potential and the limitation of exon skipping.

Exon skipping
Exon skipping induced by antisense oligonucleotides (AOs) modulates dystrophin pre-mRNA splicing process and restores the reading frame of DMD [34]. The reading frame is converted from out-of-frame to in-frame, which generates the protein that is short but functional, instead of truncated and presumably degradable dystrophin, and is considered to be capable of converting DMD to a milder phenotype [35].
There is a mutation hotspot in the DMD encoded by exons 43-45, and skipping of exon 51 can theoretically restore the largest subset of patients; therefore AOs targeting exon 51 were the first to be developed clinically [36]. There are several AO chemistries that have been investigated of which 2'O-methyl-phosphorothioate oligonucleotide (2'OMePS) and phosphorodiamidate morpholino oligomer (PMO) are the two major drug candidates that are currently being evaluated in advanced phase of clinical trials.

Phosphorodiamidate morpholino oligomer (PMO)
The PMO chemistry is a charge-neutral compound, and have reduced off-target effects, along with less immunoreactivity [37]. Phase 3 clinical trial is now being conducted for exon 51 skipping, along with phase 1 or 2 study for exon 45 and 53 skipping (https://clinicaltrials.gov/).

Others
The chemistries that have been investigated other than 2'OMePS and PMOs are  [34,37]. LNA-modified AOs display better mismatch discrimination and high resistance against nucleases; tcDNA show enhanced target-binding affinity and improved nuclease resistance [34]; PPMOS, Pip-PMOs, vPMOs contain cell-penetrating moieties [37]. Viral vectors and synthetic vectors (liposomes, cationic peptides and polymers, protein complexes) and covalent attachment such as antibodies, peptides, lipids, carbohydrates, growth factors and vitamins have been tried to overcome the poor target-specific delivery [34] and relatively unstable skipping efficiency of AOs, which are the current hurdles of exon skipping strategy.

Drugs which improve the underlying dystrophic condition
Glucocorticoids including;

Why it is important to do this review
Currently there is no curative therapy for DMD patients. Exon skipping is a therapeutic approach that have been investigated for decades which now have become the strategy at the most advanced phase of clinical trials. Review of the results from clinical studies is crucial for understanding the efficacy and the limitation of exon skipping.

OBJECTIVES
To assess whether exon skipping can positively change the clinical course of patients with DMD.

METHODS
The systematic review of the literature will comply with the PRISMA statement.

Types of studies
We will include double-blind, randomized controlled trials (RCTs) that investigate the effect of exon skipping in patients with DMD. We will also include the first phase of cross-over controlled trials.
Comparison will be done between: the effects of exon skipping and placebo, both groups without glucocorticoid therapy; the effects of exon skipping and placebo, both groups with continuous glucocorticoid therapy; the effects of exon skipping and placebo, both groups with intermittent glucocorticoid therapy.

Types of participants
All patients, including children and adults of all ages, gender, race, out and inpatients, who are confirmed to have out-of-frame DMD mutations that are identified by the authors as correctable by exon skipping are eligible for the review. The genetic analysis must be confirmed by any of the published or author approved methodologies that evaluate all DMD exons, including multiplex ligation-dependent probe, comparative genomic hybridization, single condition amplification/internal primer analysis, and target resequencing.
Studies with participants who are receiving steroid therapies are eligible for this meta-analysis, but the 13 participants must be on steroid before the first administration of exon skipping drug.

Types of interventions
Exon skipping chemistries that will be reviewed are:  Both out and inpatient administration.
 Placebo include mannitol, phosphate-buffered saline (PBS), or any other placebo if identified.
 Steroid therapies will be accepted if both treated and control groups are receiving steroid.
14 We plan to analyze data for each type of intervention separately.

Type of outcome measures
The outcomes listed here are not eligibility criteria for this review, but are outcomes of interest within whichever studies are included.

Primary outcome
We will assess the data of change in six-minute walk test (6MWT) distance from the baseline obtained between 20-35 weeks after the treatment initiation as the primary outcome, and will refer it as "data of week 24". If there are multiple data collected within the above designated weeks, than the data obtained at the week closest to week 24 will be selected for the review.
Six-minute walk test (6MWT) has been used to assess functional capacity in patients with heart and lung related diseases, and it was the primary endpoint in a study that supported regulatory approval of drug for primary pulmonary hypertension [41]. More recently, its relevance to neuromuscular diseases including DMD has been established [42].
For the participants who became non-ambulatory during the study, the 6MWT data will be recorded as 0.

Searching other resources
We will review conference proceedings for non-published studies. We will screen bibliographies of identified manuscripts for studies not identified by the search. For large clinical trials sponsored by pharmaceutical companies known to the group, but does not appear in any of the electronically searched sites, the information will be retrieved from the pharamceutical companies' websites or will be directly contacted. We will also contact the researchers who are conducting RCTs of exon skipping for information on other exon skipping RCTs.

Data collection and analysis
Two authors will independently review the titles and abstracts identified from the register and will determine for eligibility. Two authors will obtain the full text of all potentially relevant studies for independent assessment. Two authors will independently decide which trials fit the inclusion criteria.
Authors will resolve disagreements about inclusion criteria by discussion, to reach a consensus. If persistent, the disagreement will be resolved by a third author.The authorships of the studies will not be blinded prior to the assessment.

Selection of studies
We will select only randomized controlled trials, and cross-over trials for inclusion. In the Discussion, we will review open studies, longitudinal observational studies and individual case reports but will only discuss studies in which the diagnosis, intervention, pre-treatment and post-treatment states are adequately described and in whom follow up for at least six months is available.

Data extraction and management
Two review authors will independently extract data onto pre-agreed data extraction forms. We will resolve disagreements by discussion with the other authors. One author will enter data into the Cochrane statistical software RevMan (Review Manager (RevMan) [Computer program]. Version 5.3. Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014.) and a second author will check data entry. We will contact trial authors directly for any missing data.

Assessment of risk of bias in included studies
Two authors will independently assess included studies for risk of bias using pre-agreed criteria, described in the Cochrane Handbook for Systematic Reviews of Interventions (http://handbook.cochrane.org/). We will resolve disagreements by discussion with the other authors. We will grade each aspect for risk of bias as high risk ('No'), uncertain ('Unclear') or low risk ('Yes') for the following domains: • sequence generation; • allocation concealment; • blinding of participants; • blinding of outcome assessors; • incomplete outcome data; • selective outcome reporting; and • other sources of bias.

Measures of treatment effect
Statistical methods used to measure treatment effects will be in accordance with The Cochrane Handbook for Systematic Reviews of Interventions (http://handbook.cochrane.org/).
We will analyze dichotomous data as risk ratios (RR) and continuous data as mean difference, or standardized mean difference for results across studies with outcomes that are conceptually the same but measured in different ways. We will report these measures of effect with 95% confidence intervals (CI).
We will undertake meta-analyses, using RevMan, only where this is meaningful, that is if the treatments, participants and the underlying clinical question are similar enough for pooling to make sense.

Unit of analysis issues
Where a single trial includes multiple trial arms, we will include only the relevant arms. If we combine two comparisons (e.g. drug A versus placebo and drug B versus placebo) in the same meta-analysis, we will halve the control group to avoid double-counting. For each participant there may be multiple observations for the same outcome.

Dealing with missing data
If necessary, we will attempt to contact trial authors for missing data including numbers of dropouts and deaths and whether or not an intention-to-treat analysis was performed.

Assessment of heterogeneity
We will assess clinical heterogeneity by judging, qualitatively, the differences between studies regarding the participants, therapies, and reporting of important study outcomes.
We will statistically test heterogeneity of intervention effects among trials using the standard Chi 2 statistic (P value) and the Higgins I 2 statistic expressed as a percentage. We will take P values of less than 0.05 as evidence of heterogeneity. We will interpret I 2 for heterogeneity as follows (http://handbook.cochrane.org/): • 0% to 40%: may not be important; • 30% to 60%: may represent moderate heterogeneity; • 50% to 90%: may represent substantial heterogeneity; • 75% to 100%: considerable heterogeneity.
If we identify substantial unexplained heterogeneity we will report it and explore possible causes by pre-specified subgroup analysis.

Assessment of reporting biases
To detect the presence of publication bias, we will construct a funnel plot using Revman, if there are a reasonable number of studies (at least 10 in the same meta-analysis). We will use Begg's and Egger's tests to verify the bias [47,48].

Data synthesis
If there is no substantial or considerable heterogeneity, we will synthesize the data in a meta-analysis using RevMan. We will perform random-effects models for comparison purposes and use the most appropriate, depending upon the degree of heterogeneity.

Subgroup analysis and investigation of heterogeneity
If there are sufficient data, we plan to undertake the following subgroup analyses using the outcome: Within each group we will use the I 2 statistic for heterogeneity and if its value is greater than 50% we will scrutinize the trials and forest plots for differences to explain the heterogeneity. If we find no explanation, we will repeat the analysis using a random-effects model.

'Summary of findings' table
We will create a 'Summary of findings' table using the following outcomes:  Change in dystrophin gene or protein expression; and  Adverse events.
We will use the five GRADE considerations (study limitations, consistency of effect, imprecision, indirectness and publication bias) to assess the quality of a body of evidence (studies that contribute data for the prespecified outcomes). We will use methods and recommendations described in Section 8.5 and Chapter 12 of the Cochrane Handbook for Systematic Reviews of Interventions [49] using GRADEpro software. We will justify all decisions to down-or upgrade the quality of studies using foot-notes and we will make comments to aid readers' understanding of the review where necessary.

Sensitivity analysis
We will perform a sensitivity analysis to determine whether conclusions are robust by undertaking both fixed-effect and random effect meta-analysis. We will perform sensitivity analyses to assess the effect of including studies at high risk of bias on the change in 6MWT, and by repeating the meta-analysis excluding any studies at high risk of bias.

ACKNOWLEDGEMENTS
The work will be supported by The Clinical Research Program for Child Health and Development, AMED (No.27300101).

AMENDMENTS FROM THE PROTOCOL
Upon data extraction and management, the entered data into Revman software was double checked by a single author.
We planned to create a 'Summary of findings'