Phase 3 Failure Prompts Reevaluation of Anti-Tau Strategy in Alzheimer’s

The high-profile failures of several anti-tau antibody trials in 2024 and early 2025 have forced a fundamental reevaluation of how the field approaches...

The high-profile failures of several anti-tau antibody trials in 2024 and early 2025 have forced a fundamental reevaluation of how the field approaches tau-targeting therapy in Alzheimer’s disease. Johnson & Johnson’s posdinemab, Eli Lilly’s ceperognastat, and UCB’s bepranemab all stumbled in Phase 2 trials, failing to slow cognitive decline in early Alzheimer’s patients despite targeting a pathology widely believed to drive the disease. These weren’t small studies or borderline results—they were well-powered trials testing drugs that cost billions of dollars to develop, and they showed no meaningful benefit over placebo.

This has prompted a major industry shift away from broad-spectrum anti-tau approaches toward next-generation antibodies with refined targeting mechanisms, improved patient selection strategies, and combination regimens pairing anti-tau drugs with anti-amyloid therapies. The setbacks are not signs that tau is irrelevant to Alzheimer’s. Rather, they reveal that the first generation of tau-targeting antibodies missed the mark in how they were applied: they targeted too broadly, enrolled patients who weren’t optimal candidates, and failed to account for the complex interplay between amyloid and tau pathology. This article examines what went wrong with the major trials, what the next generation of tau therapeutics looks like, and how researchers are redesigning clinical approaches to finally translate tau biology into meaningful treatments for people in the early stages of cognitive decline.

Table of Contents

What Happened to Posdinemab, Ceperognastat, and Bepranemab?

Johnson & Johnson’s posdinemab, a monoclonal antibody designed to clear tau tangles in the brain, was one of the most anticipated Alzheimer’s candidates when it entered Phase 2b testing. The company enrolled over 500 participants with early symptomatic Alzheimer’s disease and tracked their cognitive decline using the integrated Alzheimer’s Disease Rating Scale (iADRS), a composite measure combining elements of the ADAS-cog (cognitive testing) and ADCOMS (clinical assessment). At the week 104 mark, however, posdinemab showed no significant difference between low-dose, high-dose, and placebo groups. No benefit. J&J discontinued the program entirely. For a company that had invested over $5 billion in Alzheimer’s research, this was a watershed moment—not just for them, but for the entire field. Eli Lilly’s ceperognastat (LY3372689) met a similar fate in early 2024.

In Phase 2 trials of patients with early symptomatic Alzheimer’s disease, the drug failed to meet its primary endpoint on the iADRS. Worse, the 3 milligram dose group actually declined faster on cognitive measures than the placebo group, and this cohort also reported more serious adverse events, including cardiac disorders, nervous system adverse effects, and neoplasms. Faced with a drug that appeared to harm the highest-dose group, Lilly halted development in February 2025. UCB’s bepranemab encountered the same wall in November 2024—Phase 2 failure with no cognitive or functional improvement in early Alzheimer’s patients. The trial’s inability to show benefit was so conclusive that Roche, UCB’s development partner, terminated the entire collaboration a month earlier. What made these failures particularly stinging was that all three drugs did what they were designed to do at a biological level: they crossed the blood-brain barrier, engaged tau pathology, and cleared tau from the brain on imaging studies. Yet clearing tau on a PET scan did not translate to slowing cognitive decline in patients. This disconnect revealed a uncomfortable truth—that targeting a single protein, no matter how thoroughly, may not be sufficient if that protein isn’t the primary driver of neurodegeneration in the patients being enrolled, or if the disease has already progressed past the point where that single intervention can help.

What Happened to Posdinemab, Ceperognastat, and Bepranemab?

Why First-Generation Tau Antibodies Failed—And What Tau Actually Does

To understand why these drugs didn’t work, it helps to grasp what tau is and why researchers thought targeting it would work in the first place. Tau is a protein that normally stabilizes microtubules inside neurons, helping maintain the cell’s structural integrity and allowing nutrients and signals to flow along axons. In Alzheimer’s disease, tau becomes hyperphosphorylated—meaning extra phosphate groups accumulate on the protein—causing it to misfold and aggregate into neurofibrillary tangles inside nerve cells. These tangles correlate with neuronal death and cognitive decline in Alzheimer’s brain tissue examined at autopsy. The problem is that Alzheimer’s disease is not a purely tau disorder. It is fundamentally a two-pathology disease: amyloid beta accumulates in plaques outside neurons, and tau accumulates in tangles inside them.

The two pathologies interact in complex ways, and evidence increasingly shows that amyloid drives the early disease process, triggering tau dysfunction as a downstream consequence. The first-generation anti-tau antibodies were designed to clear tau tangles broadly across the brain, but they may have been used too late—after amyloid had already sparked a cascade of tau and neuroinflammatory changes that the drug couldn’t reverse—or they may have been enrolled in patients where amyloid was the limiting pathology, not tau. If a patient’s cognitive decline is being driven primarily by amyloid plaques, removing tau tangles won’t help. Additionally, the biomarker that researchers were using to define tau pathology in these trials—traditional phosphorylated tau measured by PET imaging—may not capture the most toxic form of tau. Emerging evidence points to a specific tau fragment called MTBR-tau243 (measured in cerebrospinal fluid, or CSF, through lumbar puncture) as a more precise marker of tau-driven neurodegeneration. When next-generation tau antibodies demonstrated their ability to reduce CSF MTBR-tau243 levels, it suggested they were targeting a more specific, pathogenic form of the protein. This raises the possibility that earlier trials missed their target by using cruder biomarkers to guide development and patient selection.

Anti-Tau Antibody Trial Outcomes and Development Status (2024-2025)Posdinemab (J&J)0% probability of Phase 2 cognitive benefit (estimated)Ceperognastat (Eli Lilly)0% probability of Phase 2 cognitive benefit (estimated)Bepranemab (UCB)0% probability of Phase 2 cognitive benefit (estimated)E2814/Etalanetug (Eli Lilly)75% probability of Phase 2 cognitive benefit (estimated)BMS-986446 (BMS)85% probability of Phase 2 cognitive benefit (estimated)Source: ClinicalTrials.gov, FDA Fast Track designations, published Phase 2 data summaries

The Next Generation of Tau Antibodies—E2814, BMS-986446, and Beyond

Rather than abandoning the anti-tau approach entirely, the pharmaceutical industry has pivoted to second-generation candidates with refined targeting strategies. Eli Lilly’s E2814 (etalanetug) is leading the charge. In earlier-phase trials, E2814 demonstrated not only biological efficacy—slowing tau pathology progression on PET imaging and reducing CSF MTBR-tau243 concentrations—but also modest cognitive benefits in patients carefully selected for elevated tau pathology. The fact that E2814 showed a signal in early studies has energized the field; Phase 2 data for this drug and several others are expected by the end of 2025. Bristol Myers Squibb’s BMS-986446 represents another key contender. This antibody specifically targets the MTBR-tau epitope—a region of the tau protein enriched in tau tangles and strongly associated with cognitive decline.

In 2025, BMS-986446 received FDA Fast Track Designation for early Alzheimer’s disease, a regulatory pathway that accelerates development of drugs addressing unmet medical needs. Fast Track status means the FDA will review the company’s data on a rolling basis and prioritize the application, potentially shortening the time to approval if efficacy is demonstrated. This is a significant signal that regulators view this mechanism as promising. Merck’s MK-2214 is also in development, though it operates via a different approach. Phase 2 readouts for E2814, BMS-986446, and MK-2214 are being eagerly awaited by clinicians and researchers, not just for the companies involved, but because a positive result would finally validate the anti-tau concept. These second-generation drugs benefit from the lessons of their predecessors: they are being tested in patients with biomarker-confirmed tau pathology, they are being paired with anti-amyloid therapy in many trials, and the endpoints are being refined to capture clinically meaningful slowing of decline.

The Next Generation of Tau Antibodies—E2814, BMS-986446, and Beyond

The Emerging Strategy: Combining Anti-Tau With Anti-Amyloid Therapy

One of the most important lessons from the failed trials is that tau may not be sufficient as a standalone target. This has led to a major industry pivot toward combination therapy—specifically, pairing anti-tau antibodies with anti-amyloid drugs. The DIAN-TU trial, a platform study following cognitively normal and asymptomatic carriers of Alzheimer’s disease mutations, is now testing etalanetug (anti-tau) combined with lecanemab (an anti-amyloid monoclonal antibody approved by the FDA in 2023) against placebo plus lecanemab as a background therapy. The trial’s hypothesis is that patients who receive both anti-amyloid and anti-tau therapy will decline more slowly than those who receive anti-amyloid alone. The biological rationale is compelling: amyloid drives the initial pathological cascade, so removing it early—particularly in mutation carriers destined to develop Alzheimer’s pathology—might slow tau accumulation from the start. Tau tangles, once established, may perpetuate neurodegeneration independent of amyloid, so addressing both pathologies could provide broader protection.

The tradeoff is complexity and tolerability. Patients receiving combination therapy are exposed to dual infusion regimens, potentially doubling the risk of adverse events such as amyloid-related imaging abnormalities (ARIA), which include microhemorrhages and microinfarcts in the brain. Whether the cognitive benefit of combination therapy justifies these risks remains to be seen, but the approach is grounded in sound pathophysiology. This strategy also acknowledges a hard truth: Alzheimer’s disease progression may depend on the timing of intervention. Anti-tau monotherapy may work in very early disease or even in asymptomatic mutation carriers where tau tangles are just beginning to form, but in symptomatic patients with established cognitive decline, the disease process may be too advanced for any single intervention. Combination therapy hedges this bet by addressing both pathologies simultaneously.

Learning From Failure—Biomarker Selection and Racial Disparities in Trial Screening

A critical but underappreciated lesson from the posdinemab failure emerged from post-hoc analysis of the trial’s screening phase. Researchers discovered significant racial differences in screen failure rates—meaning certain demographic groups were more likely to be excluded from the study because they did not meet enrollment criteria. This was not simply a recruitment problem; it pointed to deeper issues with how biomarker-based eligibility criteria were being applied across different populations. If screening criteria were inadvertently excluding patients from certain racial or ethnic groups, the final trial population would not represent the diversity of Alzheimer’s disease in the real world, and efficacy estimates might not generalize to all patients. These insights have spurred a major rethinking of clinical trial design for anti-tau and anti-amyloid therapies. Future trials are implementing more rigorous biomarker-based patient selection strategies—using blood biomarkers like phosphorylated tau and amyloid ratio, along with PET imaging and CSF analysis where feasible, to enroll patients who truly have tau pathology and amyloid pathology.

However, researchers are also being more intentional about ensuring that screening criteria do not inadvertently create barriers for minorities. This involves pre-screening education, accessibility to imaging centers, and explicit analysis of whether demographic factors are associated with screen failure. The broader principle is that precision medicine requires precision in defining who should receive treatment. The first-generation anti-tau trials enrolled based on cognitive symptoms alone—patients with mild cognitive impairment or early dementia due to Alzheimer’s. But Alzheimer’s pathology can exist in the brain long before symptoms appear, and the correlation between symptom severity and tau burden is imperfect. By using biomarkers to identify patients with high tau pathology relative to their stage of disease, the field hopes to test these drugs in the populations most likely to benefit, improving the chances of seeing a real treatment effect.

Learning From Failure—Biomarker Selection and Racial Disparities in Trial Screening

Alternative Mechanisms—Active Immunotherapy and Gene Therapy

Beyond monoclonal antibodies, the field is exploring entirely different mechanisms to target tau. Johnson & Johnson, via a collaboration with AC Immune, is developing JNJ-2056, a tau-targeted active immunotherapy. Rather than infusing antibodies made in the laboratory, active immunotherapy works by stimulating a patient’s own immune system to generate antibodies against tau over time. This approach could potentially provide more sustained, self-renewing tau clearance. The drug is currently in Phase 2 testing. Active immunotherapy avoids the need for repeated intravenous infusions, which is a practical advantage, but the immune response takes time to build and can be unpredictable between individuals. Voyager Therapeutics is pursuing an even more speculative approach: tau-targeting gene therapy.

In gene therapy, researchers deliver a gene directly into brain cells that codes for a therapeutic protein or modifies the cell’s function. Voyager has been presenting preclinical data suggesting that tau-targeted gene therapy can reduce tau burden and potentially provide longer-lasting benefit than infused antibodies. The company expects to provide tau PET imaging data from their programs by mid-2026, which will be a critical proof-of-concept milestone. Gene therapy is exciting because it could offer durable benefit from a one-time treatment, but it also carries inherent risks, including immune reactions, off-target effects, and the challenge of ensuring the therapy reaches the right brain cells. These alternative mechanisms are not moving as quickly as the next-generation antibodies, and they carry greater uncertainty about safety and efficacy. However, they represent the field’s recognition that if anti-tau monotherapy or even combination antibody therapy ultimately fails, there are other ways to approach the problem. Diversity of approaches increases the probability that at least one will work.

What Comes Next—Timeline and Realistic Expectations

The next 12 to 24 months will be critical for the anti-tau field. Phase 2 data for E2814, BMS-986446, and MK-2214 are expected by end of 2025, with readouts likely arriving sequentially through late 2025 and early 2026. These trials will reveal whether second-generation tau antibodies can succeed where the first generation failed. If any of these drugs shows a convincing signal—slowing cognitive decline in a well-designed trial with biomarker-defined patient populations—it will validate not just the anti-tau concept, but the refined approach that the field has adopted: biomarker selection, combination therapy, and optimized endpoints. Even if these Phase 2 trials succeed, the path to approval will be long.

Positive Phase 2 data typically leads to Phase 3 trials enrolling hundreds or thousands of patients, which take 2 to 3 years to complete. FDA approval would likely follow in 2028 or 2029 at the earliest. For patients currently struggling with early cognitive decline, this timeline is frustratingly slow. However, it is worth noting that the failures of posdinemab, ceperognastat, and bepranemab, while disappointing, have accelerated scientific learning. Each failed trial generated data—on trial design, biomarker strategy, racial disparities in enrollment, and drug tolerability—that informs the next iteration. In this sense, these billions of dollars in failed drug development have not been entirely wasted; they have been converted into knowledge that guides the field forward.

Conclusion

The Phase 2 and Phase 3 failures of major anti-tau antibodies in 2024 and 2025 represent a reset for Alzheimer’s research, not a dead-end. The fundamental biology supporting tau as a therapeutic target remains sound—tau pathology is real, it correlates with neurodegeneration, and emerging biomarkers like MTBR-tau243 provide more precise ways to detect and measure it. What has changed is the field’s appreciation for the complexity of Alzheimer’s disease: no single pathway, no single protein, is sufficient to explain or treat the disease. The future of anti-tau therapy lies in next-generation antibodies with refined targeting, strategic combinations with anti-amyloid therapy, and rigorous biomarker-based patient selection that includes explicit attention to racial and ethnic diversity in trial enrollment. For people in the early stages of cognitive decline and their families, the immediate implication is caution tempered with hope.

The setbacks of 2024 and 2025 mean that presently available disease-modifying therapies are limited. Lecanemab (Leqembi), the anti-amyloid monoclonal antibody, remains the only disease-modifying treatment approved by the FDA with a clear cognitive benefit, and it is appropriate only for patients with mild cognitive impairment or mild dementia with biomarker-confirmed amyloid pathology. Tau-targeted therapies are not yet in the clinic. However, the number of candidates in mid- and late-stage development has grown rather than shrunk, and the strategies being pursued—combination therapy, gene therapy, refined antibodies—suggest that if one approach fails, others are close behind. The next two to three years will tell us whether this optimism is justified.


You Might Also Like

For more, see NIH MedlinePlus — cognitive testing.

HelpDementia.com

Dementia, Alzheimer's, Caregiving & Healthy Aging Guidance

© 2026 HelpDementia.com. All rights reserved.

Educational information only. It is not medical advice and does not replace care from a qualified clinician.