modality · intermediate · partly written
siRNA
An siRNA is a short double-stranded RNA that guides the cell's own machinery to cut one chosen messenger RNA, so the protein is not made; it acts on RNA, not on the protein or the gene.
Also called: small interfering RNA, RNAi therapeutic
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 one paragraph
An siRNA is a short RNA duplex that makes the cell cut one chosen messenger RNA, so the protein is never built. The drug supplies a sequence; the cell's own enzyme does the cutting and is reused, so one dose can act for weeks to months. Everything hard is delivery. Patisiran slowed nerve disease; revusiran, aimed at the same target, was stopped after excess deaths. 123
Reported, not yet replicated
His guess more
He said: "siRNAs, which also do that. I'm not fully sure." (that is, intervene at the RNA level). 4
Amit's hypothesis
What the evidence says about his guess more
His guess is right: an siRNA acts on messenger RNA. The patisiran label says the drug causes degradation of TTR mRNA through RNA interference. It does not bind the protein and it does not edit DNA; it stops the protein being made. What he may not have in mind is that the cutting is done by the cell's own enzyme complex, which is reused, and that the real difficulty is getting the RNA into the right cells. How it differs from an antisense oligonucleotide was not sourced here; see that record. 516
Reported, not yet replicated
The two delivery tools more
Lipid nanoparticles carry patisiran, given by intravenous infusion every 3 weeks; they stay near neutral in blood, use blood apolipoprotein E as an address to the liver, and need premedication. GalNAc, a three-armed sugar attached to the siRNA, lets liver cells grab it through their asialoglycoprotein receptor; it is given under the skin. Both go to the liver. A claim that nearly all approved siRNAs are liver-targeted was not sourced. 789105
Reported, not yet replicated
Worked and failed on one target more
Patisiran and revusiran both target transthyretin mRNA. In APOLLO (225 patients) patisiran changed a neuropathy score by -6.0 versus +28.0 on placebo at 18 months. In ENDEAVOUR 18 of 140 revusiran patients and 2 of 66 placebo patients died on treatment and dosing was stopped; TTR lowering did not differ between those who died and those who lived, and no mechanism was found. Hitting the target and helping patients are separate claims. The populations (nerve disease, advanced heart disease) and the delivery formats differ, so the pair is not a clean comparison. 23
Reported, not yet replicated
Chemistry and sequence rules deep
Which strand loads into the silencing complex depends on the relative stability of the strands' 5' ends, so designers make the guide the winner. Sugars carrying 2'-fluoro and 2'-O-methyl groups at chosen positions improved potency and duration in mice and non-human primates, the exposure data indicating that the gain was mostly metabolic stability. Details of the cutting chemistry, off-target effects through the guide's seed region and the quantified inefficiency of endosomal escape were not sourced for this record. 611
Reported, not yet replicated
Where the evidence ends deep
Delivery beyond the liver has no clear success in the sources opened. Trial results here are surrogate or composite measures in the case of the LDL and neuropathy scores. Animal durability (rodent silencing with weekly dosing for over 9 months in the founding GalNAc paper) is not a human durability claim; human durability comes from the dosing schedules above. 91213
Reported, not yet replicated
What it is
A short double-stranded RNA, 21 letters per strand in the founding work, that is loaded into a protein complex (RISC, built around Argonaute). One strand, the guide, is kept; the other, the passenger, is discarded. The guide finds a messenger RNA with the matching sequence and the complex cuts it, then lets go and repeats. Duplexes of this size were shown to silence genes in human cell lines in 2001. The strand that is kept is not random: the stability of the base pairs at each strand's 5' end decides it. As a drug, the molecule is large and heavily charged (the patisiran label gives a molecular weight around 14,300 Da) compared with a small molecule. 165
Reported, not yet replicated
Intervenes at
Same as small molecule
The development questions are the same: is there a target, does the drug engage it, does a measure of the disease move, is it safe in the people who need it. In both drugs here the target is a protein-causing gene product in the liver, and the decisive evidence is a randomised trial on a clinical measure, not on target engagement alone. 23
Our inference
Changes downstream
The active molecule is a sequence, so the design question moves from what binds the protein to which sequence is cut efficiently and which chemical modifications make it stable. The 2018 modification study found the gain in potency and duration in mice and monkeys was mostly metabolic stability. Potency here means how few milligrams per kilogram are needed, set by stability and uptake rather than binding strength. Delivery becomes the central problem: the same target gave a success with one delivery format and a stopped trial with another, though the population also differed. Measuring success means measuring target mRNA and protein in the target tissue, plus a clinical measure. 1123
Our inference
Delivery
Naked RNA is destroyed in blood and cannot cross membranes. Two routes are documented. (1) Lipid nanoparticle with an ionisable lipid, given intravenously: the lipid redesign in the 2010 paper gave silencing at 0.01 mg/kg in rodents and 0.1 mg/kg in non-human primates; apolipoprotein E acts as an endogenous address label for the liver; patients get premedication against infusion reactions. See concepts/lipid-nanoparticles. (2) GalNAc conjugate: a triantennary sugar the asialoglycoprotein receptor on liver cells pulls inside, given as a plain subcutaneous injection; the founding paper reported ED50 about 1 mg/kg after a single dose in vivo, and conjugate activity held in preclinical models with over 50 percent less receptor. Both routes lead to the liver; whether either reaches other organs was not sourced here. 789101314
Reported, not yet replicated
Persistence
It does not change DNA, so the effect is reversible and runs out when the drug is cleared (our inference from the mechanism). Dosing intervals are long compared with plasma half-lives: patisiran has a label half-life of 3.2 days yet is given every 3 weeks; inclisiran is given at day 1, day 90 and then every 6 months, and in the ORION-10 and ORION-11 trials LDL cholesterol was down 52.3 and 49.9 percent at day 510 versus placebo. The control is stopping injections. Why one dose lasts months (an endosomal depot, a long-lived silencing complex) was not established from a source opened here. LDL cholesterol is a surrogate; the European summary says there is no direct evidence yet that inclisiran reduces heart attacks or stroke. 512131
Reported, not yet replicated
Examples
Teach this
In one sentence
An siRNA is a short RNA that tells the cell's own enzyme to cut one messenger RNA, so the protein is not made, and the hard part is getting it into the right cells. 165
Reported, not yet replicated
The five-minute version
A short double-stranded RNA loads into a protein complex that keeps one strand as a guide. The guide finds the matching messenger RNA, the complex cuts it and moves on, so one drug molecule can act many times. The RNA is large and charged, so it needs a carrier: a lipid particle by infusion for patisiran, or a GalNAc sugar tag under the skin that liver cells pull in. Patisiran slowed nerve damage in a 225-patient trial; revusiran, aimed at the same target, was stopped for excess deaths with no mechanism found. 16239
Reported, not yet replicated
What a listener will ask
Does it change the gene?
Why do doses last so long?
Is it the same as an antisense drug?
Both act on RNA, but this record does not compare them from a source; see the antisense record. 1
Our inference
The common wrong idea
Because an siRNA hits the target mRNA precisely, the main problem is designing the right sequence.
The same target gave a trial win and a trial stopped for deaths; delivery format, dose and the population matter as much as the sequence. 23
Reported, not yet replicated
The one citation to hand over: 2
Still open here
- How does siRNA differ from an antisense oligonucleotide in mechanism, delivery and durability? No source for the comparison was opened here.
- Why does one dose act for weeks to months: an endosomal depot, a long-lived silencing complex, or both? No primary source was opened.
- Can siRNA be delivered to organs other than the liver? No clear success was sourced.
- How large are the off-target effects through the seed region in patients? No usable source was opened.
- Does LDL lowering by inclisiran reduce heart attack and stroke? The European summary says there is no direct evidence yet.
Linked from
- Clinical
- Dose response
- Hit discovery
- Mechanism-confirmation assays
- Medicinal chemistry: potency, selectivity, ADMET
- Lipid nanoparticles
- RNA interference
- Patisiran (Onpattro)
- Revusiran
- Antisense oligonucleotide (ASO)
- Gene therapy
- mRNA medicines and vaccines
- Why did revusiran, an siRNA against transthyretin, have more on-treatment deaths than placebo?
Sources
- Elbashir SM, et al.. Duplexes of 21-nucleotide RNAs mediate RNA interference in cultured mammalian cells. Nature (2001). paper; opened 2026-11-17 all records citing this
- Adams D, et al.. Patisiran, an RNAi Therapeutic, for Hereditary Transthyretin Amyloidosis. N Engl J Med (2018). paper; opened 2026-11-17 all records citing this
- Judge DP, et al.. Phase 3 Multicenter Study of Revusiran in Patients with Hereditary Transthyretin-Mediated (hATTR) Amyloidosis with Cardiomyopathy (ENDEAVOUR). Cardiovasc Drugs Ther (2020). paper; 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.
- Onpattro (patisiran) lipid complex injection US prescribing information. Alnylam Pharmaceuticals (2025). regulatory, not peer reviewed; opened 2026-11-17 all records citing this
- Schwarz DS, et al.. Asymmetry in the assembly of the RNAi enzyme complex. Cell (2003). paper; opened 2026-11-17 all records citing this
- Semple SC, et al.. Rational design of cationic lipids for siRNA delivery. Nat Biotechnol (2010). paper; opened 2026-11-17 all records citing this
- Akinc A, Querbes W, De S, Qin J, et al.. Targeted delivery of RNAi therapeutics with endogenous and exogenous ligand-based mechanisms. Mol Ther (2010). paper; opened 2026-11-17 all records citing this
- Nair JK, et al.. Multivalent N-acetylgalactosamine-conjugated siRNA localizes in hepatocytes and elicits robust RNAi-mediated gene silencing. J Am Chem Soc (2014). paper; opened 2026-11-17 all records citing this
- Willoughby JLS, et al.. Evaluation of GalNAc-siRNA conjugate activity in pre-clinical animal models with reduced asialoglycoprotein receptor expression. Mol Ther (2018). paper; opened 2026-11-17 all records citing this
- Foster DJ, et al.. Advanced siRNA designs further improve in vivo performance of GalNAc-siRNA conjugates. Mol Ther (2018). paper; opened 2026-11-17 all records citing this
- Ray KK, et al.. Two Phase 3 Trials of Inclisiran in Patients with Elevated LDL Cholesterol (ORION-10 and ORION-11). N Engl J Med (2020). paper; opened 2026-11-17 all records citing this
- Leqvio: EPAR summary page. European Medicines Agency (2026). regulatory, not peer reviewed; opened 2026-11-17 all records citing this
- Onpattro: EPAR summary page. European Medicines Agency (2026). regulatory, not peer reviewed; opened 2026-11-17 all records citing this
- FDA approves ONPATTRO (patisiran), announced 10 August 2018. Alnylam Pharmaceuticals investor relations (2018). organisation-page, not peer reviewed; opened 2026-11-17 all records citing this
Where it sits on the map
History of this record (6)
This record has been changed 6 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
- partly written
- Last updated
- 2026-11-19
- Record
- content/modalities/sirna.json