Therapeutic antisense oligonucleotides (ASOs) have made a splash in ALS, and neurology as a whole, over the past several years. Potential therapies for SOD1-based ALS and Huntington’s disease are currently being tested in the clinic (December 2017 news). In December 2016, the FDA approved nusinursen, developed by Ionis Pharmaceuticals in collaboration with Biogen, as a treatment for spinal muscular atrophy (see November 2016 news and and Finkel et al., 2017). And, two other ASOs, Ionis’ inotersen and Alnylam Pharmaceutical’s patisiran, are under regulatory review for hereditary transthyretin amyloidosis (see May 2018 news).
Experts discussed the potential of ASO treatments, as well as challenges in developing and implementing them, at the 2018 annual meeting of the American Academy of Neurology (AAN2018), held April 21-27 in Los Angeles.
“Antisense oligonucleotides are really bringing a new wave of enthusiasm to the [neurology] field,” commented John Day of Stanford University in California, who co-chaired a session on ASOs in the clinic.
By binding to their target RNAs, ASOs can either shut off expression of undesirable genes, upregulate ones that are needed, or modulate splicing to adjust protein expression for better neuronal health, noted Frank Bennett of Ionis Pharmaceuticals in Carlsbad, California.
In ALS, the term ‘antisense’ carries additional significance. The most common genetic cause for the disease is lengthy hexanucleotide expansions in the C9orf72 gene (see Sep 2011 news). These can be transcribed on either the sense or antisense strand into RNAs, which form inclusions (see Oct 2013 news, Nov 2013 news), and translated into dipeptides, which aggregate (see Feb 2013 news and Feb 2013 news). But it’s not yet clear which of these molecules contribute to disease, and how.
“A critical, critical issue is, what’s the role of the antisense strand?” noted Robert Brown of the University of Massachusetts Medical School in Worcester. And that has implications for the application of antisense therapeutics to C9-based ALS, which Brown and others are pursuing. “That’s the $64,000 question in the C9 field: Are we going to have to blunt both strands?”
Sense and Antisensibility in C9 ALS
Early results presented at AAN2018 suggest, maybe not. In collaboration with Ionis and working in the laboratory of Jeffrey Rothstein at Johns Hopkins University in Baltimore, Lindsey Hayes and Alyssa Coyne screened a variety of potential ASOs in patient -derived motor neurons. They identified a pair specific for either the antisense or sense portion of the C9orf72 gene. Each specifically targeted splice forms toting the repeat, but left other transcripts alone, so they didn’t affect overall levels of C9orf72 protein.
Hayes, who presented the findings, had predicted that ASOs directed at the C9orf72 sense strand would fix the bulk of cellular defects. ASOs towards the antisense C9 RNA, she thought, would add a bit more benefit.
For starters, the group focused on nuclear trafficking (see Aug 2015 news). C9orf72 expansions are known to alter the makeup of the nuclear envelope, disrupting nucleocytoplasmic transport. The researchers isolated the nuclei of ASO-treated ALS motor neurons and visualized them by super-resolution microscopy. In cells treated with the sense C9orf72 ASO, the nuclear proteins Nup50, Nup98 and POM121—mislocalized in untreated cells—returned to the nuclear pore. But, unexpectedly, the same thing happened in neurons treated with the antisense-directed ASO.
The researchers also tested localization of Ran, a crucial factor that helps shuttle proteins and RNAs through nuclear pores. In C9orf72 ALS motor neurons, Ran mislocalizes to the cytoplasm. But treatment with either ASO completely restored its nuclear location.
The data suggest that targeting either the sense or antisense C9 repeat-rich RNAs, alone, could potentially help reduce motor neuron toxicity in ALS. “This is still really surprising,” Hayes told the ALS Research Forum. “We’re still repeating and repeating, because we don’t really believe it.”
Perhaps, she speculated, expanded repeat RNAs or dipeptide repeat proteins produced from both strands contribute to the nuclear defects, but there’s some threshold at which their levels become toxic. Removing either the sense or antisense C9orf72 transcripts might, theoretically, drop the total below that threshold.
Hayes,however, said that it is too early to say whether either or both ASOs would be needed to treat the disease. There are plenty of experiments to do. The group is still working on whether the sense and antisense ASOs restore gene expression and protein profiles in motor neurons. In parallel, others are treating C9orf72 model mice with different ASOs to determine their impact in vivo.
Which ASOs are Best?
There will also be other questions to answer. One is the type of ASOs to use. Though the basic structure of an ASO is fairly simple—a short string of nucleotides homologous to the target mRNA—these oligonucleotides can be unstable. “Natural RNA and DNA make very poor drugs,” Bennett explained. Modifications of the phosphate backbone, or the bases themselves, can improve stability, tolerability, binding specificity, and targeting to cells or tissues. It is just this kind of approach that Ionis is using to develop antisense therapies for ALS and other diseases.
Brown, however, is particularly interested in the molecular structure of ASOs, and how that affects impacts efficacy (see Iwamoto et al., 2017). Most ASOs produced are a mixture of 500,000 slightly different stereoisomers, he explained. The nucleotide sequence is the same, but the orientation of individual bonds within the modified phosphate backbone varies.
To address this challenge, Brown is working with Wave Life Sciences of Cambridge, Massachusetts, which generates stereopure versions of ASOs, with all the molecular orientations tightly controlled. This enhances potency because it only uses the most effective stereoisomers. (Iwamoto et al., 2017)
To target C9orf72 repeats, the researchers first screened a handful of mixed-isoform C9orf72 sense strand ASOs. They selected one, WVE-ASO3, which reduced C9orf72 repeat-containing transcripts by 70% in a cell-free assay. When the team tested various stereopure versions of ASO3, they hit upon one conformation that further diminished levels of these potentially toxic RNAs by 90%.
To evaluate its potential as a therapy for ALS, Brown’s team tested this stereopure ASO in two different mouse models overexpressing C9orf72 repeats (O’Rourke et al., 2015; Peters et al., 2015). They injected the ASO into the brain at 1 day of age, and again a week later. By the time the mice reached 8 weeks, repeat-rich C9orf72 mRNA dropped to less than 40% of normal levels. Other C9orf72 transcripts, lacking the repeats, were unaffected. The expression of full-length C9orf72 protein was also normal.
The treatment also slashed the numbers of repeat-based RNA foci and the levels of glycine-proline dipeptides made by the repeat sequences, suggesting that it might reduce motor neuron toxicity, too. However, since the model mice exhibit no motor phenotype, it wasn’t possible to determine if the ASO would improve symptoms.
Brown’s team is currently working on further toxicology studies, and planning a clinical trial. The study could begin as early as the fourth quarter of 2018.
SOD1 ASOs: In The Offing?
While C9orf72 is an up-and-coming target for ALS therapy, ASOs targeting another ALS gene, SOD1, have been in the works for years. Tim Miller of Washington University in St. Louis, who received the 2018 Sheila Essey Award for ALS Research at AAN2018, led much of this work. “Tim has really laid the ground for ASOs in human patients,” commented Hayes.
In 2006, Miller reported that ASOs slowed disease progression and extended survival by about 10 days, in a rat model of SOD1 ALS. This suggested that the approach might be of benefit in people with the disease (Smith et al., 2006).
The approach is safe and tolerable in people with ALS, according to a phase 1 clinical trial (see May 2013 news, Miller et al., 2013). However, the researchers felt they could develop a more potent ASO, and went back to the lab before pursuing further human studies.
The new version is about 10 times more potent than the earlier SOD1 ASO, Bennett said. As Miller reported at the meeting, it is more effective in reducing SOD1 levels in mouse and rat models, and also improves survival.
In nonhuman primates, the new ASO also significantly reduced production of the enzyme in the brain and spinal cord. SOD1 mRNA levels dropped by about 75% within three days, while it took closer to a month to see a change in overall enzyme levels. This can be observed in CSF samples, Miller said, suggesting that the levels of SOD1 in the CSF could provide a biomarker for target engagement during trials. (For C9orf72 therapy, the presence of dipeptides in the CSF could alternatively be used, Brown said; see April 2017 news.)
However, preclinical studies indicate that this approach takes time to noticeably cut SOD1 protein levels. That’s because plenty of the enzyme, synthesized before treatment, persists in the CSF. To get a read on when synthesis of new SOD1 stops, Miller’s team treated animals with carbon-13, which is incorporated into new proteins. After just 30 days of ASO treatment, the carbon-13-labeled, fresh SOD1 disappeared, though older, unlabeled SOD1 remained. It’s possible to do the same labeling in people, Miller told The ALS Research Forum, which would allow clinicians to check the ASO’s action earlier. (Crisp et al., 2015)
A placebo-controlled phase 1/2 study of this new SOD1 ASO is ongoing. Miller is optimistic the treatment will reduce SOD1 protein levels in people with ALS, and, he said, “cautiously optimistic” that it will alleviate symptoms, too.
It Won’t Be Easy
Despite the potential of antisense therapy for ALS, physicians using the ASO nusinersen for spinal muscular atrophy (SMA) cautioned that this type of treatment is far from a simple prescription.
“This really is a new situation, and we are learning a lot, on an almost day by day basis,” said Day.
For example, at which stage will people likely benefit, and with which form of the disease? Nusinersen has only been tested in children. Day described one adult he’d treated. The man reported some improvement, such as to his voice strength and ability to move a computer mouse, but could this be due to the placebo effect?
Richard Finkel of Orlando, Florida, who co-chaired the session on ASOs, also noted that it’s difficult to determine dosages. The recommended schedule for nusinersen is an initial four doses, then treatment every four months. But some patients might require a dose every three months, while others may need to be treated only every 6 to 12 months. “It may not be one size fits all,” Finkel said.
Finkel also said that starting treatment before symptoms, when possible, seems to be “optimal.” Yet some parents may hesitate to initiate treatment for their children that early, especially since the penetrance and severity of SMA can vary. In terms of ALS, while SOD1 mutations virtually guarantee disease, C9orf72 repeats do not, so it’s possible carriers could face similar decisions.
Of course, those are challenges ALS specialists would love to wrestle with. And yet, even if these C9orf72 and SOD1 ASOs reach the clinic, they would only apply to those who’ve inherited these specific genetic changes. Clinicians said they’d like to order personalized ASOs, for any mutation people with ALS or other conditions their patients possess.
“We’re years away from that,” said Bennett. For now, the identification of safe and effective ASOs requires a costly screening process. But Bennett hopes that in about a decade, with better knowledge of how ASOs work, it might be possible to order personalized therapies—perhaps even from a local pharmacy.
Brown, for his part, was upbeat. He thinks some treatment—be it ASOs, gene therapy, or small molecules that affect gene expression—will work to stop ALS, if provided early in disease course.
“The age of biologics is here,” he told the ALS Research Forum. “And it is phenomenally exciting.”
Finkel RS, Mercuri E, Darras BT, Connolly AM, Kuntz NL, Kirschner J, Chiriboga CA, Saito K, Servais L, Tizzano E, Topaloglu H, Tulinius M, Montes J, Glanzman AM, Bishop K, Zhong ZJ, Gheuens S, Bennett CF, Schneider E, Farwell W, De Vivo DC, ENDEAR Study Group. Nusinersen versus sham control in infantile-onset spinal muscular atrophy. N Engl J Med. 2017 Nov 2;377(18):1723-1732. [PubMed]
Iwamoto N, Butler DCD, Svrzikapa N, Mohapatra S, Zlatev I, Sah DWY, Meena, Standley SM, Lu G, Apponi LH, Frank-Kamenetsky M, Zhang JJ, Vargeese C, Verdine GL. Control of phosphorothioate stereochemistry substantially increases the efficacy of antisense oligonucleotides. Nat Biotechnol. 2017 Sep;35(9):845-851. [PubMed]
O’Rourke JG, Bogdanik L, Muhammad AKMG, Gendron TF, Kim KJ, Austin A, Cady J, Liu EY, Zarrow J, Grant S, Ho R, Bell S, Carmona S, Simpkinson M, Lall D, Wu K, Daughrity L, Dickson DW, Harms MB, Petrucelli L, Lee EB, Lutz CM, Baloh RH. C9orf72 BAC transgenic mice display typical pathologic features of ALS/FTD. Neuron. 2015 Dec 2;88(5):892-901. [PubMed]
Peters OM, Cabrera GT, Tran H, Gendron TF, McKeon JE, Metterville J, Weiss A, Wightman N, Salameh J, Kim J, Sun H, Boylan KB, Dickson D, Kennedy Z, Lin Z, Zhang YJ, Daughrity L, Jung C, Gao FB, Sapp PC, Horvitz HR, Bosco DA, Brown SP, de Jong P, Petrucelli L, Mueller C, Brown RH JR. Human C9ORF72 hexanucleotide expansion reproduces RNA foci and dipeptide repeat proteins but not neurodegeneration in BAC transgenic mice. Neuron. 2015 Dec 2;88(5):902-909. [PubMed]
Smith RA, Miller TM, Yamanaka K, Monia BP, Condon TP, Hung G, Lobsiger CS, Ward CM, McAlonis-Downes M, Wei H, Wancewicz EV, Bennett CF, Cleveland DW. Antisense oligonucleotide therapy for neurodegenerative disease. J Clin Invest. 2006 Aug;116(8):2290-6. [PubMed]
Miller TM, Pestronk A, Devid W, Rothstein J, Simpson E, Appel SH, Andres PL, Mahoney K, Allred P, Alexander K, Ostrow LW, Schoenfeld D, Macklin EA, Norris DA, Manousakis G, Crisp M, Smith R, Bennett CF, Bishop KM, Cudkowicz ME. An antisense oligonucleotide against SOD1 delivered intrathecally for patients with SOD1 familial amyotrophic lateral sclerosis: a phase 1, randomised, first-in-man study. Lancet Neurol. 2013 May;12(5):435-42. [PubMed]
Crisp MJ, Maweuenyega KG, Patterson BW, Reddy NC, Chott R, Self WK, Weihl CC, Jockel-Balsarotti J, Varadhachary AS, Bucelli RC, Yarasheski KE, Bateman RJ, Miller TM. In vivo kinetic approach reveals slow SOD1 turnover in the CNS. J Clin Invest. 2015 Jul 1;125(7):2772-80. [PubMed]
Ling KK, Jackson M, Alkam D, Liu D, Allaire N, Sun C, Kiaei M, McCampbell A, Rigo F. Antisense-mediated reduction of EphA4 in the adult CNS does not improve the function of mice with amyotrophic lateral sclerosis. Neurobiol Dis. 2018 Jun;114:174-183. [PubMed].
Rinaldi C, Wood MJA. Antisense oligonucleotides: the next frontier for treatment of neurological disorders. Nat Rev Neurol. 2018 Jan;14(1):9-21. [PubMed].
Schoch KM, Miller TM. Antisense Oligonucleotides: Translation from Mouse Models to Human Neurodegenerative Diseases. Neuron. 2017 Jun 21;94(6):1056-1070. [PubMed].