Imagine the heartbreak of watching a loved one fade away due to Alzheimer's disease, their memories slipping like sand through fingers. While amyloid-clearing therapies have grabbed headlines, the relentless march of tau protein might hold the key to truly slowing this devastating condition. But here's where it gets controversial: despite recent failures in Phase 2 trials, a fresh wave of tau-targeting antibodies and therapies is emerging, challenging the status quo and sparking debates on what's next for Alzheimer's treatment.
Dated December 23, 2025, this report dives into highlights from the Clinical Trials on Alzheimer's Disease (CTAD) conference held in San Diego earlier that month. Anti-amyloid antibodies—both those already approved and innovative 'shuttled' versions gaining ground—dominated discussions, as covered in related stories from the meeting (parts nine, eleven, twelve, and thirteen of the series). Speakers openly admitted, however, that merely removing amyloid from the brain won't fully halt the advancement of symptomatic Alzheimer's disease (AD). For many experts, the spotlight now shifts to tau, particularly its harmful extracellular 'seeded' form, as a promising new frontier.
- Emerging antibodies are designed to prevent the spread of seeded tau.
- Fresh research links the location of tau tangles to impairments in specific areas of thinking and memory.
- Initial human trials for three antibodies targeting various tau sections indicate they are generally safe.
- Advanced labeling techniques reveal target interaction in a small study of the tau-targeted antisense oligonucleotide (ASO) NIO752.
One speaker shared findings showing that as tau tangles invade new brain areas progressively, the cognitive functions tied to those regions begin to deteriorate. These observations are correlational—they don't prove that spreading tau directly causes these brain malfunctions—but some researchers interpret them as evidence that mobile tau is indeed the villain. Antibodies offer a potential shield against this toxic advance, yet two candidates, bepranemab and posdinemab, recently fell short in Phase 2 trials. Undaunted, the scientific community pushes forward with new options. At CTAD, data from first-in-human trials of three tau antibodies, all focused on halting seeded tau propagation, were unveiled. They appeared safe overall, though, like other standard antibodies, only a minuscule fraction penetrates the cerebrospinal fluid (CSF).
Others are adopting a broader tactic, using antisense oligonucleotides (ASOs) to suppress tau production entirely. In a Phase 1 study of NIO752, a tau-ASO, stable isotope labeling kinetics help measure which tau forms are most affected.
After a string of failures with antibodies targeting the N-terminal end of tau, attention has pivoted to its central region. This midsection contains key areas that drive tau clumping, including the proline-rich and microtubule-binding regions, often abbreviated as MTBRs (as discussed in April 2018 conference coverage). Last year's CTAD showcased bepranemab, an antibody against the MTBR, as the pioneer in showing tangle reduction on tau-PET scans (November 2024 conference news). Although it didn't improve cognition for the entire Phase 2 group, it seemed to modestly slow decline in ApoE4 non-carriers who entered with low tau levels—a familiar tale of overall trial disappointment masking potential in a subgroup.
At this year's CTAD, Saori Shimizu from UCB presented tau-PET results from the open-label extension of that trial, known as TOGETHER. The initial placebo-controlled phase involved 466 participants randomized to monthly infusions of 45 or 90 mg/kg bepranemab, or placebo, over 80 weeks. Then, 387 of them joined the 48-week extension, with treatment groups sticking to their doses and the former placebo arm split to receive either 45 or 90 mg/kg bepranemab.
Shimizu detailed tau-PET data from both phases. Predictably, those switched from placebo started the extension with higher tau buildup, which then stabilized. For prior bepranemab recipients, tau levels remained steady compared to baseline. However, wide-ranging error bars at each checkpoint highlighted significant variability in scans, and to some attendees' frustration, p-values weren't computed for the extension, leaving uncertainty about statistical significance.
Just a week before CTAD, Johnson & Johnson discontinued its Phase 2 trial of posdinemab in early AD patients after an interim review showed no cognitive gains (company press release). This antibody binds to tau's proline-rich region and has demonstrated in lab and animal models that it can block seeded tau spread.
Though no full trial outcomes for Auτonomy were shared at CTAD yet, J&J's Xingjian Zhang leveraged extensive tau-PET data to explore how tau pathology's spread correlates with deficits in particular cognitive areas. In San Diego, Zhang revealed insights from analyzing correlations between tau tangle locations and scores on five domains from the Repeatable Battery for Assessment of Neuropsychological Status (RBANS). By mapping tau-cognition links voxel by voxel, he discovered alignments with known brain functions: tangles in the medial temporal lobe tied to delayed memory issues, while those in the left temporal, occipital, frontal, and frontoparietal regions linked to poorer performance in language, visual-spatial skills, immediate memory, and attention.
Zhang extended this to the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset. Since ADNI uses the ADAS-Cog13— which doesn't subdivide into domains—J&J consulted experts to map each ADAS-Cog13 question to RBANS-like categories, creating an 'ADAS-CogRB.' Testing this on Auτonomy participants (who took both tests) showed consistency with RBANS mappings. In ADNI, similar regional ties emerged between tau tangles and domain-specific struggles, implying that as tangles overtake a brain area, the cognitive tasks it supports weaken.
Zhang viewed this as backing the idea of extracellular tau seeds as a treatable target. Still, these links could mean tangles are just markers, not drivers, of regional brain dysfunction. Zhang omitted discussion of J&J's decision to shelve posdinemab.
And this is the part most people miss: despite these setbacks, many experts cling to the logic of targeting extracellular tau seeds. CTAD spotlighted Phase 1 results for three such antibodies to bolster that case.
Leading off was BMS-986446. Developed by Prothena (as PRX005) before Bristol Myers Squibb took over, this antibody attaches to the R1, R2, and R3 repeats in tau's MTBR. It had already passed safety checks in healthy volunteers via a single-dose study. In San Diego, David Walling from CenExel in Salt Lake City shared data from a small multi-dose trial. (Walling is an investigator at a CenExel site in Los Alamitos, California.) The study included 10 healthy volunteers and 12 with confirmed early AD.
Participants were assigned 3:1 to three monthly infusions of BMS-986446 or placebo. For all active recipients, it proved safe and tolerable. Out of 17 total adverse events in 13 people—three on placebo and 10 on the drug—13 were mild, four moderate. These included headaches, drowsiness, and joint discomfort. No serious issues, deaths, or discontinuations occurred, and no one developed antibodies against BMS-986446. Like other monoclonals, CSF levels were just 0.2% of plasma levels.
A Phase 2 trial for early AD is now in progress.
Next came ADEL-Y01, an IgG1 antibody against acetylated tau, specifically at lysine-280 in the VQIINK motif of its second MTBR—a form believed to fuel seeded clumping. Jointly developed by ADEL, Inc. and Osotec, Inc. in South Korea, it's under further evaluation at CenExel.
In San Diego, ADEL's Min-Seok Kim presented results from a completed single-dose trial testing 2.5, 7.5, 20, 50, or 100 mg/kg ADEL-Y01 or placebo in 40 healthy volunteers. All doses were safe and well-tolerated, with mostly mild or moderate side effects unrelated to the drug. No anti-drug antibodies appeared. The half-life ranged from two to three weeks, with CSF-to-serum ratios of 0.017% one day post-injection and 0.14% after two weeks.
An ongoing multi-dose study in people with biomarker-positive mild cognitive impairment (MCI) due to AD or mild dementia involves 12 weeks of three monthly infusions at 30 mg/kg or 60 mg/kg ADEL-Y01, or placebo. Primary goals are safety and tolerability, with secondary measures including pharmacokinetics, CSF levels, MMSE scores, CDR-SB, and exploratory fluid biomarkers.
CTAD also featured first-in-human data for VY7523, an IgG4 antibody from Voyager Therapeutics in Lexington, Massachusetts, aimed at the C-terminus of pathological tau. It was created by immunizing mice with paired helical filaments from human brains. Voyager's Elena Ratti presented single-dose trial results across six doses of VY7523 or placebo in 48 healthy volunteers. No major safety concerns arose. The 17 adverse events were all mild to moderate, with no infusion reactions. Five—headache, nausea, dizziness, and two chest pains—were possibly drug-related. One person showed anti-drug antibodies. Half-life was 22-29 days, with a CSF-to-serum ratio of 0.3%. Ratti noted that a multi-dose study in early AD has launched.
One audience member questioned the assumption that CSF presence means brain exposure, pointing out that antibodies enter CSF via the choroid plexus, not directly through brain tissue. This echoed other CTAD presentations on TfR-shuttled antibodies, which spread across brain parenchyma while traditional ones stay near vessels and in CSF (December 2025 conference news).
Ratti countered with preclinical animal data suggesting VY7523 does reach the brain, stating, 'What will be crucial is seeing if this translates to efficacy, such as changes in tau-PET signals, in upcoming studies.'
If VY7523 truly penetrates the central nervous system and influences tau-PET, we'll only know after multi-dose and Phase 2 trials wrap up.
Finally, tracking tau reduction more directly, Ross Paterson from University College London introduced a method to assess brain target engagement early. He shared interim Phase 1b data on NIO752, a tau ASO. This multi-site effort involves UCL, the Dominantly Inherited Alzheimer’s Network Trial Unit (DIAN-TU) at Washington University in St. Louis, and Novartis, creators of NIO752 designed to curb tau expression.
Paterson explained that tracking tau changes in fluids like CSF or plasma is tough because of the huge intracellular tau stores versus dilute, varied forms outside cells with slow turnover. 'This leads to a disconnect between neuronal activity and what we can measure,' he said.
To overcome this, they employed stable isotope labeling kinetics (SILK), infusing 13C-leucine—a heavier stable isotope—to label newly made tau. Mass spectrometry then tracks the 'age' of tau fragments in blood or CSF post-infusion, revealing how NIO752 impacts production without old tau pools muddying the waters.
The NIO-SILK trial enrolls five sporadic AD and five autosomal-dominant AD patients, randomized 3:2 to intrathecal NIO752 or placebo, followed 48 hours later by 16 hours of 13C-leucine IV. CSF is sampled three times over 85 days. At day 124, all get an open-label NIO752 dose, monitored until day 208. Paterson showed data from the first five enrollees plus 10 external AD controls tested at WashU.
Controls' CSF showed a gradual increase in labeled-to-unlabeled tau ratio post-infusion, slope indicating translation rate. In the blinded NIO-SILK interim data—three on NIO752, two on placebo—a suggestive trend hinted at slower tau production in three versus the other two.
Future analyses will use SILK to examine NIO752's effects on tau isoforms like hyperphosphorylated forms and fragments such as eMTBR-243.
The trial demands a lot—multiple infusions and lumbar punctures—but participants are committed, with no serious issues or dropouts yet. Paterson believes SILK will outperform static tau measures for detecting early expression changes in NIO752 and similar therapies.
Could SILK scale to bigger trials? Unlikely, due to invasiveness, but it might aid early-phase, dose-finding studies with dozens, not hundreds, of volunteers.
—Jessica Shugart
But here's the burning question: Do you think tau-targeting therapies are the future of Alzheimer's treatment, or are we chasing the wrong target? Share your thoughts in the comments—do these setbacks make you doubt the approach, or inspire hope for breakthroughs like SILK? And what if correlations between tangles and cognition mean tau is just a symptom, not the cause—does that change your view on prioritizing amyloid versus tau? We'd love to hear your take!
References
News Citations
- Amyloid Immunotherapy Affects Multiple Alzheimer’s Pathologies (https://www.alzforum.org/news/conference-coverage/amyloid-immunotherapy-affects-multiple-alzheimers-pathologies) 18 Dec 2025
- From CTAD: For Brain Shuttles, It’s All About Distribution (https://www.alzforum.org/news/community-news/ctad-brain-shuttles-its-all-about-distribution) 19 Dec 2025
- Four Years After Conception, ALZ-NET Registry Is Growing (https://www.alzforum.org/news/conference-coverage/four-years-after-conception-alz-net-registry-growing) 19 Dec 2025
- Leqembi in Asia: ARIA Is Low, Infusion Reactions Are Common (https://www.alzforum.org/news/conference-coverage/leqembi-asia-aria-low-infusion-reactions-are-common) 19 Dec 2025
- To Block Tau’s Proteopathic Spread, Antibody Must Attack its Mid-Region (https://www.alzforum.org/news/conference-coverage/block-taus-proteopathic-spread-antibody-must-attack-its-mid-region) 5 Apr 2018
- Finally, Therapeutic Antibodies Start to Reduce Tangles (https://www.alzforum.org/news/conference-coverage/finally-therapeutic-antibodies-start-reduce-tangles) 14 Nov 2024
Therapeutics Citations
- Bepranemab (https://www.alzforum.org/therapeutics/bepranemab)
- Posdinemab (https://www.alzforum.org/therapeutics/posdinemab)
- ADEL-Y01 (https://www.alzforum.org/therapeutics/adel-y01)
- VY7523 (https://www.alzforum.org/therapeutics/vy7523)
- NIO752 (https://www.alzforum.org/therapeutics/nio752)
External Citations
- press release (https://www.jnj.com/media-center/press-releases/johnson-johnson-statement-on-the-au%CF%84onomy-study)
Further Reading
No Available Further Reading
