Wrangling RNA: Antisense oligonucleotides for neurological disorders

See allHide authors and affiliations

Science Translational Medicine  25 Sep 2019:
Vol. 11, Issue 511, eaay2069
DOI: 10.1126/scitranslmed.aay2069


Effective treatment of spinal muscular atrophy with antisense oligonucleotide therapy opens the door to treating other neurological disorders with this approach.

Neurological diseases pose unique problems for medical therapy. Neurons are postmitotic, have limited capacity for regeneration, are vulnerable to age-related degeneration, and function as part of complex and precisely configured neuronal networks mostly laid down during development. Moreover, the nervous system is essentially sequestered from the systemic circulation behind the blood-brain barrier. A major challenge in advancing promising preclinical treatments is that the neurological pathways underpinning voluntary movement have intrinsic reserve capacity so that considerable undetected and potentially irreversible damage is likely to have occurred before a patient reports symptoms to a physician. Another reason why therapeutic progress in neurological disorders is slow compared to fields such as oncology is that cellular pathophysiology is less well defined. However, we are now entering an era where the first steps are being taken in precision medicine for inherited neuromuscular and other neurological diseases based on a detailed mechanistic understanding of the molecular basis of genetic mutations and how these can be manipulated.

Early treatment successes in animal models of the neuromuscular disease spinal muscular atrophy (SMA) using short single-stranded DNA–like nucleic acid compounds known as antisense oligonucleotides (ASOs) (1) have led to clinical trials and regulatory approval of ASOs for treating SMA. Altering the natural history of diseases using ASOs, which target genes through specific Watson-Crick base pairing resulting in modulation of gene splicing or expression, is now a clinical reality, with inherited neuromuscular diseases at the forefront of this new era of therapeutic intervention. In this installment of the Science Translational Medicine 10th anniversary Focus series, we discuss the hurdles that need to be overcome to expand ASO therapy beyond SMA to other neuromuscular and neurodegenerative disorders.


SMA is a neuromuscular disease caused by loss of the SMN1 gene and reduction in the widely expressed survival motor neuron (SMN) protein, resulting in selective loss of spinal cord motor neurons. Its commonest and most severe form (SMA type 1) results in lethal infantile paralysis, but all forms of SMA lead to severe disability. Because cells have an obligate requirement for SMN, a key factor in the essential cellular process of small nuclear riboprotein assembly, complete loss of SMN1 is incompatible with cell viability. The copy gene SMN2, which is differentially spliced, thus excluding exon 7 in most of the transcripts, produces small amounts of full-length SMN protein, sufficient for normal function in most cells but below the threshold required for spinal motor neurons. The number of SMN2 copies varies between individuals and correlates with disease severity in SMA. Crucially, the absolute difference in SMN protein concentrations between patients with severe or mild disease is small, suggesting that minor increases in SMN protein could have a profound clinical impact. Whereas neuroscientists are still vigorously debating which of the several functions of SMN explain selective motor neuron vulnerability, an effective treatment for SMA has arrived.

A systematic analysis of regulatory sequences in SMN2 that modulate exon 7 splicing using ASOs showed that blocking the intronic splicing silencer (ISS-N1) of the 5′ splice site led to the greatest increase in exon 7 inclusion. Although the evolutionary SMN1 duplication is not present in rodents, a series of mouse models (with human SMN2 expressed on a null Smn background) replicated the genomic architecture of the human disease and produced robust animal models of SMA, which have provided the essential confirmation that intrathecal injection of the anti–ISS-N1 ASO restored neuromuscular function (1). Subsequent work in nonhuman primates to establish dose and safety led to the first open-label and subsequent sham-controlled clinical trials of nusinersen (Spinraza) in infants with SMA, with encouraging evidence of benefit (2).

Coupled with a newborn screening program enabling ASO treatment to be started soon after birth, SMA, a previously fatal disorder, has become treatable. The development of Spinraza for SMA establishes the principle that modulating mRNA splicing can be effective therapeutically. In many ways, however, SMA is unique as a disease, and translation of ASO therapy to other neuromuscular and neurological disorders will require substantial refinements.


In contrast to the rapid and impressive development of Spinraza for SMA, progress toward effective ASO therapy for other neuromuscular disorders such as the fatal X-linked disease Duchenne muscular dystrophy (DMD) has been much slower. The DMD gene was cloned in the 1980s, and its essential function encoding the structural protein dystrophin is well understood. A wide range of deletion/duplication and nonsense mutations cause DMD by disrupting the open reading frame, resulting in the absence of full-length functional dystrophin protein. The first evidence that the effects of such mutations could be abrogated through use of splice site–modifying ASOs to correct an aberrant reading frame emerged in the mid-1990s. In the early 2000s, in vivo efficacy of such ASOs was demonstrated in the dystrophic mdx mouse model of DMD (3), in which the therapeutic benefit of a morpholino phosphorodiamidate ASO that rescued muscle dystrophin expression was shown. This ultimately led to the 2016 approval by the U.S. Food and Drug Administration (FDA) of an ASO (eteplirsen) that resulted in skipping of exon 51 in the transcript encoding dystrophin and restoration of dystrophin protein (4, 5). However, the approval of eteplirsen was not without controversy. Although the safety data were supportive, the efficacy of the intravenously administered ASO was in question, especially given the very low amount of dystrophin protein generated (<1% of normal dystrophin protein abundance).

Key differences between ASO development for DMD and SMA must be considered. Given that numerous mutations in DMD lead to disease, multiple exon-skipping ASOs targeting separate exons will be required to treat a majority of patients with DMD (e.g., only about 13% of patients would be candidates for eteplirsen treatment). Although the fold difference in SMN protein concentrations between SMA patients with severe or mild disease is relatively small, this is not true in the case of DMD, where increasing dystrophin protein to at least 10% of normal expression will likely be required for therapeutic efficacy. Further, although both diseases could be regarded as systemic in nature, the major cellular target in SMA is the spinal cord motor neuron, which can conveniently be targeted via the local intrathecal route. In the case of DMD, all skeletal muscle groups (especially those relevant to respiration) and cardiac muscle should be targeted, given that cardiorespiratory failure is the primary cause of premature death in patients with DMD. Achieving this challenging goal requires substantially higher ASO drug doses administered through a systemic route. Notably, an exon 51–skipping ASO (drisapersen) was rejected by the FDA, principally on the basis of safety concerns. There are encouraging clinical trial data regarding related exon-skipping ASOs, including Sarepta Therapeutics’s golodirsen and NS Pharma’s viltolarsen that both target exon 53, most likely reflecting improved ASO length, target sequence, and dose (compared to eteplirsen). However, accelerated approval for golodirsen was very recently declined by the FDA, highlighting the continued challenges in developing ASO therapy for DMD.


Most genetically determined neurodegenerative disorders are late-onset autosomal dominant conditions in which the gene mutation acts through altering the protein product in a way that leads to acquired toxicity. The task for therapy is therefore to antagonize the aberrant gene product, at the RNA or protein level, without driving toxicity through loss of function of the normal protein.

Amyotrophic lateral sclerosis (ALS) is an aggressive neurodegenerative disease of motor neurons in which the average survival is 2 to 3 years from onset and for which substantially disease-modifying treatments are currently lacking. About 12 to 15% of patients with ALS carry a disease-determining genetic mutation. There appear to be many biological triggers of ALS, given that mutations in more than 20 different genes have been implicated. The SOD1 and C9orf72 genes, however, account for about 60 to 70% of ALS mutations, and therefore, these two genes have become the focus for ASO therapy. Although highly expressed in the nervous system, rodent knockout experiments suggest that ablation of SOD1 expression is tolerated. This has led to the development of an ASO targeting both mutant and normal SOD1, with the primary aim of reducing the accumulation of misfolded mutant SOD1 protein. This is required because of the impracticality of delivering personalized ASO therapy for the more than 100 separate missense mutations described in ALS cases. Rodent ALS models treated by intrathecal administration of tofersen (previously IONIS-SOD1Rx), a 2′MOE gapmer (which induces RNA degradation via the intracellular enzyme RNase H), resulted in mutant SOD1 protein reduction and an extension of survival. This ASO is now in phase 3 clinical trials in patients with ALS ( NCT02623699).

A hexanucleotide expansion mutation in the first intron of the gene C9orf72 accounts for up to 10% of all ALS cases and also a substantial fraction of cases of the neurodegenerative disorder frontotemporal dementia. The mechanism of disease toxicity is still debated, but most evidence suggests that a rational therapeutic strategy is to block the production of the repeat RNA to mitigate direct RNA toxicity or the production of dipeptide repeat proteins, which arise when intronic repeat RNA is translated via a non–ATG-dependent mechanism. Because concern exists that haploinsufficiency may play a role in C9orf72-related neurodegeneration, ASOs have been designed to target the C9orf72 pre-mRNA. These ASOs reduce the repeat-containing transcript without affecting total C9orf72 protein and have shown positive effects in reducing toxicity in cellular models. These cellular models include induced pluripotent stem cell–derived motor neurons, which have been crucial due to the lack of appropriate rodent models for C9orf72 mutations (6). A clinical trial to assess the safety and toxicity of the ASO IONIS-C9Rx is now under way in patients with ALS who carry C9orf72 mutations ( NCT03626012).

Similarly, IONIS-HTTRx, a 5-10-5 2′MOE gapmer targeting the HTT gene responsible for the neurodegenerative disorder Huntington’s disease (HD), at a site distant from the CAG repeat mutation, has undergone initial clinical studies in patients with early manifest HD (7). The early signals from this clinical trial ( NCT02519036), with reduced mutant huntingtin protein in the cerebrospinal fluid, are encouraging. Initial analysis has not yet demonstrated any difference in clinical outcome related to reduced mutant huntingtin protein. A competing clinical approach in HD by Wave Life Sciences takes an allele-specific strategy in which the stereochemistry of the oligonucleotides is controlled with the aim of improved efficiency in mitigating toxicity and preserving function of the wild-type huntingtin protein.


Evidence from SMA in particular, and emerging data from other disorders, indicates that we are entering a new age of precision genetic medicine for neurological disorders, led by a maturing ASO technology (Fig. 1). Rapid progression beyond a few of the most obvious neuromuscular disease targets requires the development of next-generation ASO technologies. At present, the intrathecal route remains the most practical, as it allows local delivery to the central nervous system, limits the potential for systemic toxicity, and minimizes production costs. To achieve nervous system disease modification using a systemic route more practical for long-term use, these next-generation drugs will have to offer improved potency and safety. Advanced delivery technologies could drive ASO delivery across the blood-brain barrier, thus circumventing the need for repeated intrathecal drug administration. A potentially interesting technological advance is the advent of stereopure ASO chemistry (8), permitting chirally controlled ASO synthesis (current ASOs are typically chiral mixtures), with improvements in both potency and safety, and offering the potential of allele-specific targeting. A plethora of delivery technologies, including protein/peptide-based (9) and exosome-based nanotechnologies (10) are emerging for enhanced intracellular ASO delivery to overcome the very poor intracellular bioavailability of nucleic acid drugs. These, coupled with next-generation ASO chemistries, are likely to herald an age of much wider application of ASO medicines. Ultimately, transformation of the therapy of neuromuscular and neurodegenerative disorders necessitates presymptomatic treatment (as is now beginning for SMA), requiring appropriate early screening programs and biomarkers to guide effective treatment intervention. Although there are many technical challenges ahead, the first steps toward enabling the realization of disease-modifying therapies for currently untreatable neuromuscular and neurodegenerative diseases have been taken.

Fig. 1 Advancing ASO therapy for neurological disorders.

Future advances in ASO therapy for neuromuscular and neurodegenerative disorders will depend on both technical advances (left) and improved screening and biomarker strategies for identifying patients during the earliest stages of disease (right). Development of next-generation ASO compounds will arise from advances in oligonucleotide chemistry, including stereoselective synthesis, coupled with technologies that solve the challenges of intracellular ASO delivery and delivery across the blood-brain barrier. Last, the ability to assess the true clinical effectiveness of ASO therapies will be driven not only by presymptomatic genetic screening of neonates or at-risk family members but also by development of more robust clinical trial outcome measures, including development of digital technologies and devices that will have implications for reductions in clinical trial size, duration, and ultimately cost.



Acknowledgments: We thank T. Roberts for contributing to the figure. Competing interests: K.T. has served on scientific advisory boards for Roche, Cytokinetics, and Biogen in relation to SMA and ALS. M.J.A.W. is co-founder and non-executive director of PepGen and Evox Therapeutics and currently collaborates with Wave Life Sciences. He is also a non-executive director of Oxford University Innovation and is a named inventor on numerous patents related to oligonucleotides and nucleic acid drug delivery systems.
View Abstract

Navigate This Article