PGx7529: Exploratory Pharmacogenetic evaluation of Potential Modifiers of Duchenne Muscular Dystrophy patient response to Drisapersen

Trial description: GlaxoSmithKline (GSK) is developing drisapersen (GSK2402968, formerly PRO051), an antisense oligonucleotide directed against exon 51, for the treatment of Duchenne Muscular Dystrophy (DMD). DMD is a progressive neuromuscular disorder that is ultimately lethal and is caused by genetic variation in the dystrophin gene (DMD) resulting in the absence of the dystrophin protein. About 13% of DMD patients have one of eight different out-of-frame DMD deletions that can be put in-frame by skipping exon 51, the largest proportion of DMD patients that may be treated with the same oligonuclotide. Results of the Phase II study (DMD114117; n=53) indicated a clinically meaningful and statistically significant difference from placebo after 24 weeks of treatment with 6mg/kg/wk drisapersen.The results of the Phase II clinical study DMD114876 (n=51) were that a clinically meaningful treatment benefit over placebo in the six minute walk distance (6MWD) at Week 24 (not statistically significant at the 5% level) was observed for the 6 mg/kg/wk treatment arm. The 3 mg/kg/wk group showed no clinically relevant difference from placebo in 6MWD at Week 24. In September 2013, it was announced that the Phase III double-blind, placebo-controlled clinical study (DMD114044) of 186 boys treated with 6mg/kg drisapersen weekly injection did not meet the primary efficacy endpoint when compared to placebo. No clinically meaningful or statistically significant treatment difference in key secondary assessments of motor function were observed. Through their participation in the drisapersen clinical development program, subjects were invited to provide optional written informed consent and a saliva sample for pharmacogenetic (PGx) research. GSK proactively generated subject genotypes using DNA extracted from saliva samples to address questions that might be raised by the clinical study results. GSK aims to identify genetic predictors that may provide insight into the discordant clinical study results. None of the individual or combined clinical studies were designed or powered to address specific genetic hypotheses, so any genetic findings should be considered exploratory. The primary analysis investigating genetic predictors of change from baseline in observed and predicted 6MWD will be conducted using the DMD114044 data. Additional analysis will be conducted across studies by either combining data on the subject level accounting for study or by combining study specific effect estimates using meta-analysis. Disease progression will be investigated by combining data across treatment groups, including placebo, under the assumption of lack of efficacy of drisapersen on disease progression. Efficacy will be examined within active treatment arms. In both instances, terms for treatment may be included in the model. In addition to accounting for ancestry as an independent variable, analyses may be conducted by major ancestral subgroup depending on the sample size per subgroup. Subgroup analyses may also be performed by age groups (e.g. ≤7, >7), by DMD mutation category, or by other criteria driven by clinical data. Multiple test adjustments will not be made to account for multiple studies, endpoints or subgroup analyses. The genetic markers are assigned to tiers based on prior evidence for involvement in disease progression or efficacy. Tier 1 will be comprised solely of SPP1 rs28357094 given this has the strongest prior evidence of involvement in DMD disease progression. Tier 2 will be comprised of the remaining candidate gene markers. Multiple testing corrections for the effective (independent) number of genetic markers will be made using a Bonferroni adjustment within each tier based on maintaining an experimentwise type 1 error, α, of 0.05; α will be split among the two tiers allocating an α of 0.03 to tier 1 and 0.02 to tier 2. Based on prior information, G allele carriers will be compared to TT homozygotes for SPP1 variant rs28357094. For the remaining markers, an additive genetic model based on the number of minor alleles will be used. Effect estimates, standard errors and p-values will be displayed for results meeting the predefined statistical thresholds or for results of prior interest. A mixed model repeated measures approach will be used to examine the effect of marker genotypes on change from baseline in 6MWD and change from baseline in percent-predicted 6MWD. This analysis will be conducted using SAS software PROC MIXED.