In the phase II CELIA trial, the investigational tau-targeting antisense oligonucleotide (ASO) diranersen has shown therapeutic potential in patients with early Alzheimer’s disease (AD), yielding substantial reductions in tau burden and slowing clinical decline despite missing its primary endpoint.
“CELIA establishes proof of concept, as the first randomized phase II trial of a tau-targeting agent to show that a reduction in tau pathology may slow early AD progression,” said principal investigator Dr Catherine Mummery from the University College London, London, UK.
At week 76, an unprecedented reduction in tau pathology was seen across all diranersen dose groups, Mummery noted. “This reduction was observed across multiple brain regions.”
Diranersen-treated patients had mean reductions in CSF tau of between 50 percent to 65 percent. In the subset of those underwent tau PET scan, tau signals decreased across the medial and lateral temporal cortex, frontal cortex, parietal cortex, anterior and posterior cingulate cortex, occipital cortex, and whole brain grey matter. [Mummery C, et al, AAIC 2026]
Results for five out of six clinical endpoints favoured all doses of diranersen over placebo, with the largest treatment effects consistently observed at the low dose (60 mg), according to Mummery.
Compared with placebo, low-dose diranersen slowed clinical decline by 26 percent on the Clinical Dementia Rating-Sum of Boxes (CDR-SB), by 42 percent on the 13-task Alzheimer’s Disease Assessment Scale–Cognitive Subscale, by 50 percent on the Mini-Mental State Examination, by 23 percent on the Alzheimer’s Disease Composite Score, and by 30 percent on the modified integrated Alzheimer’s Disease Rating Scale.
For the CDR-SB subdomains, patients on low-dose diranersen had 20 percent to 42 percent less decline on the cognitive domain and 21 percent to 29 percent less decline on the functional domain.
CELIA did not meet its primary endpoint of CDR-SB dose response, given that the lowest diranersen dose showed the most consistent positive clinical effect, as Mummery pointed out.
“Based on these data, diranersen will advance to phase III development,” she said.
In a statement, Dr Maria Carrillo, Alzheimer’s Association chief science officer and medical affairs, described the data as encouraging.
“Today’s approved disease-modifying treatments target beta amyloid, and they are changing lives. But Alzheimer’s is a complex disease, and we must continue to pursue every promising research approach,” Carillo said. “For the first time in a phase II trial, researchers demonstrated robust removal of tau from the brain, and that is the kind of signal that gives the field reason for optimism.”
Mechanism and safety
Diranersen is an intrathecally administered antisense oligonucleotide that targets microtubule-associated protein tau (MAPT) RNA to reduce both intracellular and extracellular tau.
“Diranersen works upstream of tau deposition,” Mummery said. “Effectively, it’s a synthetic oligonucleotide that binds to the messenger RNA produced from the gene, leads to its degradation, and therefore reduces the production of tau across all forms of tau, including toxic elements, and across all compartments.”
Early signs of the drug’s clinical efficacy were seen in the phase Ib study, wherein diranersen demonstrated target engagement (ie, reduction in CSF tau) and impacted tau pathology (tau PET). Exploratory analyses showed numerically slower decline on clinical scales with the 60- and 115-mg doses vs an external control at week 100. [Nat Med 2023;29:1437-1447]
In terms of safety, the drug is well tolerated, according to Mummery.
Most adverse events (AEs) documented in CELIA were mild or moderate, nonserious, and did not lead to treatment discontinuation. The most common AEs were procedural pain, postlumbar puncture syndrome, and confusional state. Mummery pointed out that most episodes of confusional state occurred within 7 days of dosing and resolved within a week.
Serious AEs occurred in 13.1 percent to 23.3 percent of diranersen-treated patients, with the lowest rate observed in the low-dose group. AEs related to lumbar puncture or intrathecal administration were reported in 37.7 percent to 57.7 percent across the diranersen dose groups. Nevertheless, the high study completion (82 percent) and long-term extension rollover rates (94 percent) suggest that intrathecal administration was not a barrier to treatment adherence, as Mummery noted.
One patient in the placebo group and one in the diranersen 115 mg Q24W group died, but none of the deaths were related to treatment. There were no cases of amyloid-related imaging abnormality (ARIA) observed during the study, as anticipated.
The CELIA trial
CELIA included 416 patients (mean age 68 years, 50 percent female) who met criteria for mild cognitive impairment due to AD or mild AD dementia. All had confirmed amyloid pathology based on PET or CSF tests.
The patients were randomly assigned to receive diranersen at 60 mg every 24 weeks (n=60), 115 mg every 24 weeks (n=115), or 115 mg every 12 weeks (n=116) or placebo (n=115). Treatment was administered intrathecally over 76 weeks.
“Additional analyses are ongoing and will be presented at future congresses and in literature,” Mummery said.