modality · intermediate · complete
Antisense oligonucleotide (ASO)
An antisense oligonucleotide (ASO) is a short, chemically armoured single strand of nucleic acid that pairs with one RNA to destroy it, block it or change how it is spliced.
Also called: ASO, antisense oligo, gapmer, splice-switching oligonucleotide
This describes published medicine for readers who want to understand how it works and what happened. It is not medical advice and says nothing about what any person should take.
In brief
An antisense oligonucleotide is a short, chemically armoured strand that pairs with one RNA. Depending on its design it either recruits an enzyme that cuts the RNA, so less protein is made, or it just sits on the RNA and changes how it is spliced, so a different protein is made. It is chosen by sequence, injected, and lasts for months. 12
Established
His guess: it acts at the RNA level
His words: "Antisense oligos, which I guess intervene at the RNA level." 3
Amit's hypothesis
Checking the guess against the evidence
Correct, and the sources add a refinement. An antisense oligonucleotide acts on RNA by base-pairing with it, as the FDA label for nusinersen and the oligonucleotide chemistry review describe. But "at the RNA level" covers two quite different actions: destroying the RNA (a gapmer, via the enzyme RNase H) or only blocking a site so splicing changes (a steric blocker such as nusinersen). The same sequence logic can lower a protein or switch on a different version of it. 12
Established
Two modes: cut or block more
The deciding structural fact is whether the oligo and its target RNA look like a DNA:RNA hybrid. RNase H cleaves the RNA strand of such a hybrid; the review says RNA:RNA duplexes do not elicit it, which is why a gapmer design is needed to recruit it.
A gapmer has a central stretch of DNA-like letters (the gap) flanked by modified wings. The wings give stability and binding; the gap gives the enzyme a substrate. The 5-10-5 layout (a 20-mer) is used by mipomersen, inotersen and volanesorsen. Result: the target RNA is destroyed and the protein falls.
A steric blocker has every position modified, so the enzyme cannot engage. Nusinersen is an 18-mer of this kind. It occupies a site and changes what the splicing machinery does. Which enzyme isoform does the cutting is not named in the sources opened. 1
Reported, not yet replicated
The chemistry is the drug more
Natural nucleic acid does not survive. Two standard fixes appear on the nusinersen label: phosphorothioate links (one oxygen replaced by sulfur) and a 2'-O-methoxyethyl sugar. The review reports that phosphorothioate DNA binds plasma proteins, mainly albumin and alpha2-macroglobulin, and that oligo chemistry in general affects Toll-like-receptor binding, coagulation and complement activation (it does not tie each toxicity to one chemistry). MOE adds nuclease resistance and about 1 degree C of melting-temperature gain per residue. 12
Reported, not yet replicated
Delivery decides the design more
After intravenous dosing an all-phosphorothioate MOE gapmer had a plasma half-life of 15 to 45 minutes in five species including human. Two routes dominate.
To the brain and cord: injection into the spinal fluid, because the drug does not cross the blood-brain barrier well. In the nusinersen label the plasma trough is low compared with CSF.
To the liver: a sugar tag (triantennary GalNAc) binds a receptor on liver cells only. In mice it gave 6 to 10 fold higher potency for a MOE gapmer and about 60 fold for shorter cEt gapmers. Mouse potency is not human potency; only the abstract was read. 124
Reported, not yet replicated
How long it lasts and how it is stopped more
Nusinersen: terminal half-life 135 to 177 days in CSF and 63 to 87 days in plasma, cleared by exonucleases. The 2016 label doses months apart. The only way to stop is to stop dosing and wait for washout; this is reasoned from the half-lives. Later label revisions were not opened. This describes a label, not guidance for any person. 2
Reported, not yet replicated
One class, two fates: nusinersen and tominersen more
Nusinersen switches a splice in spinal muscular atrophy: in the randomised sham-controlled ENDEAR trial the final motor-milestone response was 37 of 73 (51 percent) versus 0 of 37, with a hazard ratio of 0.53 for death or permanent ventilation.
Tominersen aimed to lower huntingtin: CSF mutant huntingtin fell by 47 percent on the every-2-months dose, yet that arm did worse than placebo and dosing was stopped in March 2021. The cause is a hypothesis, not a finding. See the two drug records. 5678
Reported, not yet replicated
Splice switching in detail (SMA) deep
The SMN1 gene is lost or defective in spinal muscular atrophy; the near-identical SMN2 gene mostly skips exon 7. The trial protocol puts the skipping at up to about 90 percent and describes an intronic silencer (ISS-N1). A tiling screen of 31 ASOs found two that strongly raised SMN2 exon 7 inclusion in transgenic mouse liver and kidney (abstract only). The nusinersen label says it binds a sequence in the intron downstream of exon 7. The gene is turned up, not knocked down: the opposite of a gapmer. See concepts/splicing. 9102
Reported, not yet replicated
Backbone class toxicities deep
On the nusinersen label, platelets fell below the lower limit of normal in 6 of 56 treated patients (11 percent) versus 0 of 28 sham, none under 50,000, and urine protein was raised in 17 of 51 (33 percent) versus 5 of 25 (20 percent); the label asks for monitoring of both. The review links oligo chemistry to Toll-like-receptor binding, coagulation and complement activation but does not tie each effect to phosphorothioate alone. 21
Reported, not yet replicated
Where it breaks deep
A gapmer cuts any RNA with enough match, so sequence screening and checks on related transcripts are needed; splice switchers can shift unintended exons. Exposure is not effect: a big fall in the target did not guarantee benefit (tominersen). Delivery is compartment-limited. Animal potency gains are not human potency. For tominersen the more frequently dosed arm showed more of the neuron-injury marker NfL than placebo (+10 percent against +6 percent) while the protein marker fell, which separates "lower the protein" from "tolerate the drug" in the data. The evidence read here is mostly abstracts plus the 2016 label, the protocol and a sponsor-written summary. 162
Reported, not yet replicated
What it is
A short strand of nucleic acid, around 18 to 20 letters for the approved examples, built to pair with one RNA by Watson-Crick base-pairing. Natural DNA or RNA would be chewed up by nucleases, so the real drug is the sequence plus a chemical armour: the phosphate links are swapped for phosphorothioate and the sugar carries a 2'-O-methoxyethyl (MOE) group, as on the nusinersen label. 12
Established
Intervenes at
Same as small molecule
The path from disease to drug still runs through a causal target, a potency measure, safety testing and a trial against a control. The target is chosen by what the RNA looks like in sequence, not by a three-dimensional pocket, but the failure points are familiar: engagement of the target is not the same as benefit, and class toxicities appear in the clinic. 126
Our inference
Changes downstream
Potency becomes how much RNA is destroyed or how much the splicing shifts per tissue exposure, not binding in a pocket. Selectivity becomes sequence uniqueness: a cutting design (gapmer) can cut any RNA with enough match, so many sequences are screened along the target (tiling). Toxicity is partly a property of the backbone (platelet and kidney effects on the nusinersen label). Dose spacing follows how long the drug stays in the target tissue; the label dose interval of months follows from a cerebrospinal fluid half-life of 135 to 177 days. Manufacturing is chemical synthesis of a polymer rather than a small-molecule route; that part was not researched here. 1210
Our inference
Delivery
Naked ASOs are large, charged and cleared quickly from blood: an all-phosphorothioate MOE gapmer had a plasma half-life of 15 to 45 minutes across mouse, rat, dog, monkey and human. They do not cross the blood-brain barrier well, so for brain and cord disease nusinersen is injected by lumbar puncture into the cerebrospinal fluid. For the liver, a triantennary GalNAc sugar tag binds a receptor found on liver cells; in mice it improved the potency of a MOE gapmer 6 to 10 fold, about 60 fold with shorter cEt gapmers, and the tag is clipped off inside the cell. Subcutaneous absorption of that gapmer was 80 to 100 percent. 124
Reported, not yet replicated
Persistence
Months, set by the chemistry and the tissue. Nusinersen has a mean terminal half-life of 135 to 177 days in CSF and 63 to 87 days in plasma and is cleared by exonucleases, not CYP450 enzymes; the 2016 label spaces doses months apart. There is no off switch: the drug is not a gene, so effect fades by clearance after dosing stops. That last point is reasoning from the half-lives, not a sourced statement. 21
Our inference
Examples
Teach this
In one sentence
An antisense oligonucleotide is a chemically armoured strand that pairs with one RNA and either gets it cut or changes how it is spliced. 12
Established
The five-minute version
Small molecules usually grab a protein. An antisense oligonucleotide acts one step earlier, on the RNA, chosen by sequence rather than a pocket. Natural strands do not survive, so the backbone is armoured (phosphorothioate links, MOE sugars). If the design leaves a DNA centre the enzyme RNase H cuts the RNA and the protein falls; if every position is modified the oligo only blocks a site, as nusinersen does to switch on a different splice of SMN2. It cannot cross the blood-brain barrier, so for brain and cord it is injected into spinal fluid, where it lasts months (half-life 135 to 177 days for nusinersen). Nusinersen worked in a sham-controlled trial; tominersen lowered its target and still did not help. The lesson is that hitting the RNA is the easy half. 1256
Reported, not yet replicated
What a listener will ask
Is it gene therapy?
Does it always destroy the RNA?
Why is the brain delivery route an injection?
The common wrong idea
Antisense drugs work by destroying a bad RNA.
That is only one mode (the gapmer); nusinersen leaves the RNA intact and changes how it is spliced so more of a good protein is made.
The one citation to hand over: 1
Still open here
- What does ASO manufacturing, scale-up and quality control involve? It was not researched in the sources opened.
- Why did the more frequently dosed tominersen arm do worse than placebo? Hypotheses only; none is proven.
- Which isoform of RNase H does the cutting, and how general is the DNA:RNA-hybrid rule? The enzyme isoform was not named in the sources opened.
- Why does SMN2 exon 7 skipping give an unstable protein? Not stated in the sources opened.
- Which off-target RNAs matter in humans, and how are they predicted? No source opened covers this.
- How well does spinal-fluid delivery reach deep brain regions in disease? Not established in the sources opened.
- Which other antisense drugs were withdrawn or refused, and why? Mentions were not verified: the evidence for a refused exon-skipping drug and for early withdrawn drugs was not opened.
Linked from
- Clinical
- Dose response
- High-throughput screening
- Hit discovery
- Medicinal chemistry: potency, selectivity, ADMET
- RNA interference
- Nusinersen (Spinraza)
- Tominersen
- Gene therapy
- siRNA
- A trial enrolment figure that disagrees inside one sponsor's own material not yet written
- Why did tominersen make patients worse while lowering mutant huntingtin?
Sources
- Egli, M., Manoharan, M.. Chemistry, structure and function of approved oligonucleotide therapeutics. Nucleic Acids Research (2023). review; opened 2026-11-17 all records citing this
- U.S. Food and Drug Administration (label text supplied by the sponsor). SPINRAZA (nusinersen) injection, for intrathecal use: US prescribing information (initial US approval 2016, revised 12/2016). FDA Drugs@FDA label archive (2016). regulatory, not peer reviewed; opened 2026-11-17 all records citing this
- Amit's dictation of 2026-11-17: what the atlas must contain. Kept in the atlas repository; not published.
- Prakash, T. P., et al.. Targeted delivery of antisense oligonucleotides to hepatocytes using triantennary N-acetyl galactosamine improves potency 10-fold in mice. Nucleic Acids Research (2014). paper; opened 2026-11-17 all records citing this
- Finkel, R. S., et al.. Nusinersen versus sham control in infantile-onset spinal muscular atrophy (ENDEAR). New England Journal of Medicine (2017). paper; opened 2026-11-17 all records citing this
- F. Hoffmann-La Roche Ltd (sponsor). GENERATION HD1 (BN40423) final results: plain-language summary, September 2023. Roche forpatients (2023). organisation-page, not peer reviewed; opened 2026-11-17 all records citing this
- BioSpace. Ionis and Roche halt Huntington's disease trials. BioSpace (2021). news, not peer reviewed; opened 2026-11-17 all records citing this
- HDBuzz. HDBuzz explainer on the GENERATION HD1 results (article 354). HDBuzz (patient-community science news) (2023). news, not peer reviewed; opened 2026-11-17 all records citing this
- Sponsor-written trial protocol. Clinical trial protocol NCT02594124 (nusinersen extension study), full document. ClinicalTrials.gov large documents (2015). regulatory, not peer reviewed; opened 2026-11-17 all records citing this
- Hua, Y., et al.. Antisense masking of an hnRNP A1/A2 intronic splicing silencer corrects SMN2 splicing in transgenic mice. American Journal of Human Genetics (2008). paper; opened 2026-11-17 all records citing this
- Yamamoto, et al.. [Descriptive title, not as printed] Exposure-response analysis of tominersen across four studies. British Journal of Clinical Pharmacology (2025). paper; opened 2026-11-17 all records citing this
- Tabrizi, S. J., et al.. Targeting huntingtin expression in patients with Huntington disease. New England Journal of Medicine (2019). paper; opened 2026-11-17 all records citing this
Where it sits on the map
History of this record (2)
This record has been changed 2 times, on the dates below. The note kept against each change is written for the people who maintain the atlas, in their own shorthand, so it is not shown here; it is in the record itself.
- Status
- complete
- Last updated
- 2026-11-17
- Record
- content/modalities/antisense-oligonucleotide.json