In this report, valiltramiprosate 265 mg BID produced rapid, substantial, and sustained reductions in plasma p-tau217 over 78 weeks in both the MCI subgroup and overall APOE4/4 homozygote population in the Phase 3 trial. The effect was stronger in MCI than in Mild AD, where only a nonsignificant numerical trend was observed. Similar long-term reductions were seen over four years in the Phase 2 trial among APOE4 carriers and heterozygotes with MCI. These findings support our hypothesis that valiltramiprosate’s plasma biomarker effects are disease-stage dependent and mirror its clinical and volumetric imaging benefits, with greater effects in MCI than in the more advanced Mild AD population.

We interpret the reduction in plasma p-tau217 as evidence that valiltramiprosate inhibits amyloid oligomer-driven tau phosphorylation, consistent with its established molecular mechanism [17, 18]. This interpretation is further supported by a recent cellular study showing that tramiprosate, the active moiety of valiltramiprosate, modulates the amyloid cascade and reduces tau phosphorylation in a concentration-dependent manner, potentially through inhibition of GSK-3β phosphorylation [31].

In APOE4/4 MCI subjects in the Phase 3 APOLLOE4 trial, plasma p-tau217 decreased by week 26 and remained reduced through week 78, with a 36% decrease on valiltramiprosate compared with a 17% increase on placebo, yielding an approximately 53% treatment-placebo difference. The p-tau217/Aβ42 ratio showed a similar pattern, with an approximately 47% treatment-placebo difference. Plasma p-tau217 is a circulating biomarker of amyloid-induced tau hyperphosphorylation and is mechanistically downstream of amyloid oligomers [7]. By inhibiting aggregation of amyloid monomers into soluble neurotoxic Aβ oligomers [17, 18], the proximal driver of tau phosphorylation, valiltramiprosate would be expected to reduce p-tau217 as a direct marker of target engagement; the magnitude and timing of the observed reductions are consistent with this interpretation.

We previously reported the valiltramiprosate Phase 2 topline results in APOE4 carriers who received valiltramiprosate 265 mg BID over 2 years, the same dose as the ALZ-801 Phase 3 trial. These results showed significant, sustained reductions in plasma p-tau181, the primary outcome of the Phase 2 biomarker trial, along with slowed hippocampal atrophy and stabilization of RAVLT memory performance [19]. Together with the present analysis, these Phase 2 findings show sustained, directionally consistent reductions in both plasma p-tau217 and p-tau181, suggesting that this pharmacodynamic signal is reproducible across study designs and APOE4-defined patient populations. For context, plasma p-tau217 has been reported to increase by at least 20% over 4 years in untreated biomarker-positive MCI participants in the A4 Study [30]. Moreover, the greater p-tau217 reduction in MCI than in Mild AD is consistent with the Phase 3 clinical findings, in which benefits on cognition, function, and brain atrophy were greater in the prespecified MCI subgroup than in the Mild AD subgroup [21, 24].

The high AD positivity rates across both trials, based on the plasma p-tau217/Aβ42 cut-off threshold, provide important context for interpreting these biomarker findings within the broader ATN framework of AD [7, 11]. Positivity was confirmed in 94% of the overall Phase 3 population, 91% of the Phase 3 MCI subgroup, 97% of Phase 2 APOE4/4 subjects, and 96% of Phase 2 APOE3/4 subjects. Within the ATN framework, plasma p-tau217 reflects amyloid-induced tau pathology downstream of the amyloid cascade but upstream of overt neurodegeneration. Recent studies show that plasma p-tau217 is highly concordant with CSF tau biomarkers and amyloid PET in AD populations [9, 10, 12]. The high plasma-confirmed AD positivity rate in Phase 2, which enrolled subjects using CSF biomarker criteria, further validates the Fujirebio Lumipulse plasma p-tau217/Aβ42 threshold against a gold-standard CSF-enriched AD population and is consistent with evidence that plasma p-tau217 has diagnostic accuracy comparable to CSF-based biomarkers in AD [30, 32]. Together, these AD positivity enrichment rates support the plasma p-tau217/Aβ42 ratio as a robust enrollment tool for APOE4/4 homozygotes and APOE3/4 heterozygotes, with the potential to achieve trial enrichment comparable to the recent anti-amyloid antibody studies without requiring invasive CSF collection or large-scale PET imaging.

Importantly, in the Phase 3 MCI cohort, valiltramiprosate-related changes in plasma p-tau217 correlated significantly with changes in cognition, function, and brain volumetric measures, including ADAS-Cog13, CDR-SB, HV, cortical thickness, and whole brain volume. The direction of these associations was consistent across domains: greater reductions in plasma p-tau217 were associated with less cognitive decline, less functional deterioration, and reduced brain atrophy. Although not reported here, diffusion tensor imaging (DTI) metrics from MRI were previously analyzed and confirmed that the observed vMRI benefits reflected preservation of neuronal tissue rather than fluid shifts or edema [21, 24]. In the MCI group, reduction in plasma p-tau217 was also correlated with stabilization of plasma NfL, a biomarker of neuroaxonal injury. This pattern suggests that the p-tau217 pharmacodynamic response reflects attenuation of the downstream neurodegenerative cascade rather than an isolated biochemical effect. Supporting this interpretation, cross-sectional associations between baseline p-tau217 and baseline disease severity were confirmed in the overall Phase 3 MCI population, consistent with plasma p-tau217 as a biomarker of existing burden of pathology. Together, these findings indicate that plasma p-tau217 serves not only as a pharmacodynamic marker of valiltramiprosate’s effect on amyloid-induced tau pathology, but also as a potential biomarker of disease severity that tracks neurodegeneration across cognitive, functional, and structural domains in this population, consistent with its established role within the ATN framework [30].

For context, analyses of late-stage AD trials suggest that an approximately 20–30% reduction in plasma p-tau217 versus placebo may be needed for translation into clinical benefit. In Phase 2 and Phase 3 donanemab trials that demonstrated clinical efficacy, plasma p-tau217 decreased by approximately 20% from baseline in the active treatment arms and increased by approximately 10% with placebo over 78 weeks. By contrast, in two Phase 3 semaglutide trials involving approximately 4000 patients with early AD that failed to show clinical benefit, plasma p-tau217 did not differ significantly from placebo, and CSF p-tau217 was reduced by only 10% versus placebo [33]. Valiltramiprosate exceeded this apparent threshold of 20%-30%, reducing plasma p-tau217 by 36–45% from baseline in the Phase 3 and Phase 2 MCI groups, with even larger differences versus placebo. Notably regarding drug effects on plasma NfL, in Phase 3 trials of both donanemab and lecanemab in early AD, plasma NfL increased in both the active treatment and placebo arms over a similar timeframe, and the active-placebo differences were not statistically significant despite larger sample sizes [26, 27]. Also, donanemab-related effects on plasma NfL did not correlate significantly with clinical outcomes [15]. It is therefore notable that valiltramiprosate significantly stabilized plasma NfL in MCI subjects (p=0.032 vs. placebo at week 78), that correlated with p-tau217 reduction, thus providing convergent evidence of slowed axonal neurodegeneration. This placebo-controlled trial in APOE4/4 homozygotes is among the first late-stage interventional AD studies to show that both p-tau217 and NfL changes with valiltramiprosate correlated with each other and with cognitive benefit [28, 29].

Importantly, reductions in plasma p-tau217 produced by amyloid plaque-clearing antibodies have not yet been prospectively linked to clinical benefit at the individual subject level. In the Phase 2 donanemab trial, despite substantial plaque clearance and measurable reductions in plasma p-tau217, p-tau217 changes did not correlate significantly with the primary clinical outcome (iADRS; r = − 0.04, p = 0.61) or whole brain volume change (r = +0.14, p = 0.10) [15]. This pharmacodynamic-clinical dissociation is further supported by two Phase 3 gantenerumab trials, which showed robust reductions in plasma p-tau217 and other AD biomarkers without clinical benefit [34]. We propose that the mechanism driving p-tau217 reduction may determine whether these biomarker changes translate into clinical outcomes. Plaque clearance by anti-amyloid antibodies may reduce circulating p-tau217 as a downstream consequence of removing insoluble amyloid deposits, yet may not sufficiently reduce the soluble neurotoxic oligomers that drive ongoing synaptic injury and tau hyperphosphorylation. By contrast, valiltramiprosate is thought to act upstream of antibodies by directly inhibiting soluble amyloid oligomer formation, thereby reducing the proximal driver of amyloid oligomer-induced tau phosphorylation. Together, these findings suggest that associations between plasma p-tau217 changes and clinical or brain structural outcomes may be mechanism-dependent, reflecting therapies that directly attenuate the oligomeric amyloid-tau cascade rather than any treatment that lowers circulating p-tau217.

Consistent with the clinical findings, valiltramiprosate showed its strongest effects in the MCI subgroup, with clear, sustained reductions in plasma p-tau217 and significant correlations with clinical and brain volumetric outcomes. By contrast, the Mild AD subgroup showed only a numerical reduction that did not reach statistical significance. This disease-stage-specific pattern aligns with the APOLLOE4 efficacy results, in which the MCI subgroup showed the greatest cognitive and functional benefit [21]. Biologically, these findings may reflect a critical treatment window for valiltramiprosate, when amyloid oligomer-driven tau phosphorylation remains the dominant active process and neuronal reserve is still sufficient to support measurable pharmacodynamic and clinical responses. APOE4/4 homozygotes exhibit accelerated hippocampal atrophy and cortical thinning at the amnestic MCI stage, before overt cognitive decline becomes the dominant clinical feature [4], making this stage a potentially optimal window for intervention with an amyloid oligomer inhibitor. Together, these findings support a precision-development strategy for valiltramiprosate focused on APOE4/4 homozygotes and APOE4 carriers at the MCI stage, where the pharmacodynamic signal is strongest, the biomarker-clinical relationship is most pronounced, and the biological rationale is most closely aligned with the drug’s mechanism of action [35, 36].

One limitation of our report is that the statistical analyses were based on observed cases and were not adjusted for multiplicity [21]. However, observed case analyses avoid assumptions about specific covariates that may influence the magnitude of the drug effects or their temporal profile. In longitudinal plasma biomarker analyses of interventional AD trials, the use of covariates to model individual biomarker trajectories requires further study, because different fluid biomarkers, including p-tau isoforms, NfL, and amyloid monomers, likely reflect distinct disease processes with different onset and progression rates that may be influenced by APOE4 genotype, comorbidities, and other factors. Additional statistical approaches, including alternative models and partial correlations between early biomarker changes and later clinical or imaging outcomes, are an area of significant interest for future investigation. Another limitation is that the Phase 2 MCI APOE4 carrier subgroup was small because of the overall trial size; therefore, these plasma biomarker results should be interpreted in that context. A third limitation is that the Phase 2 trial was a single-arm study. The 4-year plasma p-tau217 data, while not placebo-controlled, reflect sustained longitudinal changes from baseline based an objective biomarker endpoint, and are therefore considered supportive evidence.

Because valiltramiprosate inhibits Aβ oligomer formation [17], future interventional trials of anti-amyloid agents may increasingly evaluate Aβ oligomers as disease biomarkers that reflect mechanistic target engagement for drugs targeting the amyloid aggregation pathway. Recent advances in the area of Aβ oligomers assays are promising. For example, Blömeke et al. [37, 38] quantified Aβ oligomers in plasma and CSF from subjects with early AD using surface-based fluorescence distribution analysis. Although very encouraging, this method did not clearly distinguish oligomer burden across the subject groups studied, possibly because the detected oligomer range was limited by the specificity of the capture antibody, bapineuzumab. We had previously developed a sensitive Aβ oligomer assay using IMS-MS technology to identify and characterize drug effects on specific individual oligomer species [17, 18]. Development of this investigational oligomer assay is ongoing, with efforts focused on improving sensitivity so that it can advance from controlled in vitro studies to real-time application in biological matrices [39]. This and similar assays may help detect Aβ oligomers in clinical samples, clarify their pathophysiological role in AD progression, and aid in assessment of target engagement of anti-oligomer agents in AD clinical trials.

The present findings may help advance the use of p-tau217 as both an inclusion criterion and a biomarker endpoint in future valiltramiprosate studies. The FDA approval of the Fujirebio Lumipulse p-tau217/Aβ42 assay as a diagnostic aid for AD provides a practical, validated tool for enrolling subjects with biomarker-confirmed AD. In addition, the 91–94% AD positivity rates in the Phase 3 trial show that plasma biomarker-based enrichment is feasible at scale without requiring CSF collection or amyloid PET imaging. In the valiltramiprosate-treated MCI cohort, correlations between changes in plasma p-tau217 and changes in clinical and brain structural outcomes support p-tau217 as a clinically relevant pharmacodynamic endpoint, rather than only a diagnostic marker, in trials targeting amyloid oligomer pathology. Its potential role as a surrogate outcome for clinical efficacy in MCI and AD prevention trials with anti-oligomer agents warrants further study.

In summary, valiltramiprosate produced sustained long-term reductions in plasma p-tau217. These reductions correlated with clinical and brain volumetric benefits and with plasma NfL, a biomarker of neurodegeneration. Together, these findings support the promising efficacy of valiltramiprosate in APOE4/4 homozygotes and APOE4 carriers with MCI and may help guide the design of planned confirmatory studies.