A mutant repeating DNA sequence produces a toxic RNA molecule that causes the neuromuscular disorder myotonic dystrophy type 1. An ‘antisense’ therapy that targets this RNA in cell nuclei shows promise in mice. See Letter p.111
Myotonic dystrophy type 1 is a devastating inherited disorder for which effective treatments are lacking. It is characterized by myotonia (delayed relaxation of muscles after contraction), weakness, cardiac arrhythmias, diabetes and cognitive changes. Patients carry a mutant DMPK gene that contains a greatly expanded DNA repeat sequence within a non-coding region1,2,3. When the mutant gene is expressed, it yields a messenger RNA molecule that seems to be toxic to cells such as muscle fibres4 — which means that the inheritance of a single mutant copy of DMPK from either parent is sufficient to cause disease. On page 111 of this issue, Wheeler and colleagues5 describe the successful use of antisense oligonucleotides (synthetic RNA-like fragments that bind to a target RNA) to correct the molecular and physiological features of the disease in a mouse model.
The mutation in DMPK elicits the toxicity associated with myotonic dystrophy type 1 (DM1) through at least two mechanisms4. Usually, DMPK-encoded mRNA is synthesized in the cell's nucleus and efficiently exported to the cytoplasm, where it is translated into protein (Fig. 1a). However, owing to its expanded repeat, the mutant RNA forms a hairpin-shaped structure that binds to members of the MBNL family of proteins, which regulate RNA splicing, a process by which immature mRNAs are cut up and reassembled before translation (Fig. 1b). As a result, the mutant RNA and the proteins form aggregates ('foci') in the nucleus, and MBNL activity is decreased. In a second mode of action, the mutation somehow triggers increased levels of a different splicing protein, CELF1. The opposing effects of the DMPK mutation on CELF1 and MBNL activities disrupt RNA splicing and so lead to disease4.
A potential strategy for treating DM1 is to 'silence' the mutant DMPK using antisense oligonucleotides (Fig. 1c). Because patients carry a functional copy of the gene in addition to the altered one, oligonucleotides could be designed that would bind selectively to the mutant RNA and mark it for degradation — while leaving translation of the functional RNA unaffected. This strategy can be tested in a mouse model of DM1, in which the expanded repeat from a mutant DMPK has been added to an unrelated gene, the expression of which is easy to track in muscle6. Indeed, it has been shown that direct injection of a specific type of antisense oligonucleotide (morpholino oligonucleotides) into the muscles of such mice reduced the toxicity of the mutant RNA, although systemic delivery proved inefficient7.
Wheeler et al. used 'gapmer' antisense oligonucleotides, which contain chemical modifications at their ends that make them more stable. Moreover, gapmer oligonucleotides include a central sequence that, when bound to its target RNA, promotes target cleavage by the enzyme RNase H. The authors administered gapmer oligonucleotides to DM1 model mice by subcutaneous injection, and observed a robust and sustained decrease in the concentration of the toxic RNA in muscle, even when applying relatively low doses. This decrease was associated with improvement in a wide range of disease features — for example, the authors observed loss of nuclear foci, release of MBNL proteins from nuclear aggregates, and correction of RNA splicing defects and the resultant myotonia. Astonishingly, the most effective oligonucleotides continued to confer some benefit up to a year after treatment had been discontinued.
Animal tests of RNA-directed therapies for muscle diseases such as DM1 have had limited success so far7,8,9. So, why this apparent breakthrough? It comes down to the therapy's probable site of action: the nucleus. Most mRNAs are synthesized and spliced in the nucleus, then rapidly exported to the cytoplasm. But gapmer oligonucleotides induce degradation of RNA by RNase H, which is enriched in the nucleus and almost absent from the cytoplasm. Wheeler and colleagues, and a second research group working independently9, reasoned that the nuclear retention of expanded-repeat RNAs could make them good targets for RNase H. Consistent with this idea, the authors describe how oligonucleotides designed to target RNAs that are rapidly exported to the cytoplasm were ineffective at decreasing their expression in muscle. By contrast, oligonucleotides targeting a nuclear RNA (the long non-coding RNA Malat1) demonstrated similar efficacy to that seen when targeting the expanded-repeat RNA.
What are the implications of Wheeler and colleagues' results? They inspire optimism that previous challenges faced by researchers looking at antisense oligonucleotide therapies for DM1 and other neuromuscular diseases are surmountable — although significant hurdles remain regarding safety and delivery to affected tissues other than skeletal muscle, such as the heart and brain. The authors' findings also suggest that gapmer-based strategies might be suitable for the treatment of other disorders caused by expansions of repeated DNA sequences (such as amyotrophic lateral sclerosis and frontotemporal dementia10,11), provided that the mutant RNA tends to remain in the cell's nucleus longer than the normal RNA. Furthermore, appropriately designed gapmer oligonucleotides may aid researchers in defining the functions of specific nuclear non-coding RNAs, some of which have key roles in regulating gene expression.
However, as promising as these findings are for the prospect of DM1 therapeutics, they also serve as a cautionary tale for the applications of antisense oligonucleotides. First, given our limited understanding of the roles of nuclear non-coding RNAs and the likelihood that their sensitivity to this technology is enhanced, care must be taken in oligonucleotide design to avoid potentially deleterious off-target effects. Second, developing similarly potent therapies for target mRNAs that are rapidly exported from the nucleus may require the use of oligonucleotides that do not act through RNase H. Third, therapeutic success in a mouse model is still a long way from effective application in humans. However, the path to success now seems clearly visible.
Brook, J. D. et al. Cell 68, 799–808 (1992).
Harley, H. G. et al. Nature 355, 545–546 (1992).
Buxton, J. et al. Nature 355, 547–548 (1992).
Renoux, A. J. & Todd, P. K. Prog. Neurobiol. 97, 173–189 (2012).
Wheeler, T. M. et al. Nature 488, 111–115 (2012).
Mankodi, A. et al. Science 289, 1769–1773 (2000).
Wheeler, T. M. et al. Science 325, 336–339 (2009).
Mulders, S. A. et al. Proc. Natl Acad. Sci. USA 106, 13915–13920 (2009).
Lee, J. E., Bennett, C. F. & Cooper, T. A. Proc. Natl Acad. Sci. USA 109, 4221–4226 (2012).
Renton, A. E. et al. Neuron 72, 257–268 (2011).
DeJesus-Hernandez, M. et al. Neuron 72, 245–256 (2011).
About this article
Journal of Cardiovascular Translational Research (2013)