Dystrophin is a 427 kDa protein composed of four main domains, including an N-terminal actin-binding domain, a large central rod domain containing 24 spectrin repeats, a cysteine-rich region, and a carboxy-terminal domain. Although no true complete population-based systematic assessments have been performed, most mutation surveys indicate that approximately 70% of all DMD-causing dystrophin mutations are due to single or multiexon deletions with a higher mutational frequency observed within exons 44–55, which corresponds to the rod domain of dystrophin. Such deletions alter the reading frame of dystrophin and result in a prematurely truncated protein [13
]. Given the architecture of the dystrophin protein, it is expected that for many of the DMD patients with rod deletions, restoration of dystrophin's reading frame by the targeted removal of an additional exon from the mature transcript will restore a partially functional dystrophin protein and thus provide clinical benefit. This expectation is based on studies showing that patients carrying large, in-frame deletions within the rod domain of the dystrophin protein frequently exhibit a milder clinical phenotype (referred to as Becker muscular dystrophy or BMD) [14
]. A compelling example is that some individuals with large in-frame mutations, spanning exons 45–55, remained asymptomatic until 69 years of age [16
]. These observations led to the hypothesis that the central rod domain of dystrophin was dispensable for dystrophin function. Elegant experiments from Jeff Chamberlain's group [17
] defined the critical regions of dystrophin by testing the ability of mutant dystrophin, with internally truncated rod domain deletions, to rescue the mdx
phenotype. Their investigations also demonstrated that larger deletions sometimes lead to a milder phenotype than smaller deletions [17
]. These studies, along with the recognition that natural, in vivo
`exon skipping' occurred in the mdx
] and in humans with DMD [20
], established the validity of targeting RNA splicing to restore the proper reading frame as a therapy.
Therapeutic exon skipping is now being tested in animal models of dystrophin deficiency and in human DMD trials. These studies utilize antisense oligonucleotides (AONs) to direct the lack of inclusion of targeted exons containing nonsense or frame-shift mutations into the translated mRNA. Between 2001 and 2003, the feasibility of exon skipping was demonstrated with the successful administration of oligonucleotides and induction of exon skipping in mdx
mice in vivo
]. In 2005 and 2006, successful systemic administration was accomplished in the mdx
mouse, although the efficiency was not yet to therapeutic levels [25
]. Since that time, a variety of chemistries and delivery methods have been devised and tested in the mdx
mouse and many researchers have continued to identify strategies to improve the efficiency of delivery, while keeping in mind toxicity and immunogenicity. For example, phosphorodiamidate morpholinos coupled to Arg-rich, cell-penetrating peptides effectively restored dystrophin in 96% of mdx
skeletal muscle fibers, but were less effective in cardiac muscle (58%) [27
]. Other studies using octaguanidine-coupled morpholinos demonstrated that the efficiency of delivery could be improved with this modification [26
]. Studies have administered 2-O
-methyl oligonucleotides to mice for as long as 8 months of treatment with continued phenotypic improvement apparent in the mice at 16 months of age [29
], suggesting that this approach may be tolerated for extended periods of time, an important feature for a chronic disease such as DMD.
Whereas some researchers have focused upon improving the chemistries of the AONs to improve efficacy, others have coupled the oligonucleotides to various carriers to improve delivery. Agents such as nanopolymers of polyethylene glycol and polyethyleneimine [30
] and polylactide-co-glycolic acid nanospheres [31
] and cationic core shell nanoparticles [32
] were used to deliver charged AONs (2-O
-methyl) to mdx
muscles. Although promising, all of these studies will require additional exploration of their potential toxicities.
An alternate approach to systemic antisense-based exon skipping has been proposed and tested in cell culture and the mdx
mouse, in which the AON is cloned in tandem with a modified U7 small nuclear RNA sequence and expressed from an adeno-associated virus [33
]. Although this requires a gene-therapy-like approach with its attendant problems, the possibility of a more permanent repair without the need for continued therapy is appealing. Improvements in this process have recently been published in which the AON is also linked to a short-binding sequence of heterogeneous ribonucleoprotein A1 [33
It is not clear how small antisense sequences interfere with RNA splicing as the process is complex and is influenced by numerous RNA-binding proteins and splice enhancer sequences. Devising universal therapies targeted to specific exons is further complicated by the uniqueness of each DMD mutation and the associated deletion breakpoints. Another difficulty is that the optimal specific sequence to target is not always clear. Although some oligonucleotides are effective if targeted to splice donor and acceptor sites, these motifs are not always preferred targets [34••
]. Furthermore, it will be imperative to optimize the oligonucleotides used for therapeutic intervention in the context of human cells in vitro
] or in the transgenic mouse expressing human dystrophin [35••
], prior to initiating clinical trials.
The first human clinical trials for exon skipping are focused on exon 51, because AONs that efficiently induce exon 51 skipping were identified and because of the relatively large proportion of patients for whom exon 51 skipping would generate an in-frame dystrophin transcript. Patients with specific exon deletions (e.g. Δ 47–50, Δ 48–50, Δ 49–50, Δ 50, and – Δ52) are in aggregate 13% of the DMD population and constitute the most common therapeutic targets in whom the skipping of a single exon is needed to restore reading frame. Trials are being conducted in Europe, targeting exon 51 using two different chemistries. In the Netherlands, researchers administered 2-O
-methyl AONs that hybridize to an internal sequence of exon 51 (called PRO051) into the tibialis anterior muscle of four DMD patients bearing genetic deletions that were correctable by exon 51 skipping [36
]. Biopsied, treated muscles from each patient exhibited detectable levels of dystrophin protein without adverse effects, demonstrating successful exon skipping and establishing a key landmark for proof-of-principle studies in humans. Based on these promising results, phase I/II studies using systemic administration of PRO051 via subcutaneous delivery are underway and will test the safety and efficacy of a 5-week treatment regime and 13 weeks of follow-up. In parallel, local introduction of AVI-4658, which also targets the same region of exon 51 through an alternate backbone chemistry called morpholino, into the extensor digitorum brevis muscle was tested over a year ago and unpublished results indicate that some dystrophin expression was restored in the injected muscle. Similarly, a 12-week, phase I/IIa systemic delivery clinical trial of AVI-4658 has been initiated at Imperial College London by Drs Francesco Muntoni and Katherine Bushby (unpublished observation). Prior to the study, several different oligonucleotide chemistries were tested using cultured human muscle cells and using a mouse expressing a human dystrophin gene as a model system to identify optimal oligonucleotide conditions [37
Mutational data indicate that following exon 51 skipping, the next six most common single exon skip targets are exons 45, 53, 44, 46, 52, and 50 (in that order based on frequency of the DMD mutations). Recently, two studies have demonstrated that exon skipping can also be used with complicated dystrophin mutations that lead to `pseudoexons'. Thus, it is possible that more different types of mutations than it had previously been thought may benefit from this sort of therapeutic approach [38
] and some point mutations will be amenable to this therapy. Given the large size and exon structure of DMD, there are a staggering 76 different single exons that could be therapeutically targeted in at least one observed mutation. Thus, even if exon 51-targeted therapy is successful, a tremendous amount of work is needed to develop a comprehensive approach to generate an armory of genetic mutation-targeted therapies, which will require an infrastructure to develop, design, and test each targeted therapeutic that may be used in only a single child. This challenge will make DMD a compelling experimental area for truly personalized medicine.
Notable reports of exceptionally mild (or asymptomatic) BMD in patients with large exon 45–55 deletions have led some investigators to explore the feasibility of developing a cocktail of AONs, which could be used as a single drug to treat as many as 63% of all patients with DMD [16
]. Recent animal studies have led to encouragement that this approach may be feasible. Wilton and coworkers were able to use antisense oligomers to successfully restore dystrophin's reading frame in the mdx4cv
mouse model, which is one that requires double skipping (of exons 52 and 53) to place the dystrophin transcript back in frame [41
]. In addition, Hoffman and coworkers demonstrated the feasibility of multiple exon skipping using the canine muscular dystrophy dog model (CXMD), which has a mutation at the splice site of exon 7 of the dystrophin gene [42••
]. To correct this mutation and to restore the reading frame requires skipping two exons (6 and 8) to create a fusion between exons 5 and 9. The authors used a cocktail of antisense morpholinos injected into the leg veins of the dogs and demonstrated some variable correction in all muscles tested of each dog, including the heart, but to a lesser extent. Thus, multi-exon skipping has now been successfully carried out in vivo
in both small and large animal models. These studies are very encouraging for the development of an AON cocktail that could treat a large percentage of dystrophin mutations. Cocktails of 2-O
-methyl oligonucleotides against all exons between 45 and 55 were tested in human cells in vitro
]. Unfortunately, the researchers were not able to identify a cocktail that was effective for inducing such a large deletion; thus, additional studies will be necessary before therapeutic exon skipping between exons 45 and 55 becomes a reality.
These preliminary studies demonstrate that exon skipping is a viable strategy to induce the production of dystrophin in DMD boys. However, the success of this therapeutic approach rests on overcoming the inefficiencies of exon skipping, as it is unclear whether the levels of skipped dystrophin currently being achieved will be sufficient to functionally reverse the disorder, particularly in DMD boys. Best estimates indicate that 30–60% of wild-type levels of skipped dystrophin will be required to functionally compensate for loss of dystrophin [17
]. Early trial data, though promising, indicate that even high-dose local intramuscular delivery of AON falls short of inducing such levels, yielding only 3–35% of normal dystrophin. It is anticipated that systemic delivery of AON may be even more inefficient. Therefore, it will be imperative to increase the efficacy of exon skipping to replace dystrophin to functionally relevant levels, which is being pursued by altering the oligonucleotide backbone, altering the targeted sequence, modifying attachments to the oligonucleotide sequence, increasing delivery to muscle, and identifying small molecule facilitators of exon skipping. Nonetheless, the success of the early trials with PRO051 is encouraging.