Plasma Biomarker Effects of Oral Valiltramiprosate/ALZ-801 in Early Alzheimer’s Disease from Phase 3 and Phase 2 Trials: Analysis of Core Biomarkers and Correlations with Clinical and Neuroimaging Outcomes

Abstract

Background

Alzheimer’s disease (AD) is a progressive neurodegenerative disorder with few treatment options, especially for APOE4/4 homozygotes who carry high genetic risk. Valiltramiprosate/ALZ-801 is an oral small-molecule inhibitor of amyloid-beta oligomer formation in late-stage development as a disease-modifying therapy for AD. A recent Phase 3 trial did not meet its primary clinical endpoint in the overall population, but demonstrated significant benefits in cognition, function, and brain volumetric MRI (vMRI) in the prespecified subgroup with mild cognitive impairment (MCI). Here, we report plasma p-tau217, p-tau217/Aβ42, and neurofilament light chain (NfL) results from the Phase 3 trial, along with 4-year fluid biomarker data from the Phase 2 study in APOE4 carriers with early AD.

Methods

The 78-week, APOLLOE4 Phase 3 trial randomized APOE4/4 homozygotes to placebo (n = 162) or valiltramiprosate 265 mg BID (n = 163). The Phase 2 trial was an open-label biomarker study in APOE4 carriers with early AD who received valiltramiprosate 265 mg BID (N = 84) for 104 weeks, followed by a 2-year extension on the same regimen. Plasma p-tau217 and p-tau217/Aβ42 were measured every 6 months in both trials using FDA-approved Fujirebio Lumipulse G assay. Plasma NfL was measured in the Phase 3 trial using the Simoa assay. Associations between plasma biomarkers and clinical or vMRI outcomes were assessed using Spearman’s correlation. Data are reported separately by study.

Results

Baseline plasma p-tau217/Aβ42 confirmed amyloid positivity in 94% of 294 Phase 3 APOE4/4 subjects, including 91% of those with MCI, as well as in 97% of 80 Phase 2 subjects. In the overall Phase 3 population, valiltramiprosate significantly reduced plasma p-tau217 compared with placebo at weeks 26, 52, and 78 (all p < 0.025, observed values). A similar effect was seen in the Phase 3 MCI subgroup (p < 0.05 at weeks 52 and 78), whereas the Mild AD subgroup showed only a numerical trend. In observed case analyses of MCI subjects, plasma p-tau217 decreased by 36% from baseline with valiltramiprosate and increased by 17% with placebo. In the same subgroup, the p-tau217/Aβ42 ratio decreased by 40% with valiltramiprosate and increased by 7% with placebo. Among Phase 3 MCI subjects, week 78 reductions in p-tau217 correlated significantly with improvements in ADAS-Cog13 (r = 0.276, p = 0.039), CDR-SB (r = 0.377, p = 0.005), hippocampal volume (r = − 0.353, p = 0.013), cortical thickness (r = − 0.337, p = 0.018), and whole brain volume (r = − 0.312, p = 0.029). Plasma NfL was also significantly lower with valiltramiprosate than placebo at week 78 in MCI subjects (p = 0.032). NfL stability over 78 weeks correlated significantly with improvements in ADAS-Cog13 (r = 0.29, p = 0.03) and hippocampal volume (r = − 0.282, p = 0.047), and with reductions in p-tau217 (r = 0.42, p = 0.001). In Phase 2 APOE4 carriers, valiltramiprosate-induced reductions in plasma p-tau217 were sustained over 4 years and remained significantly lower in both MCI APOE4 carriers and heterozygotes than in the Phase 3 APOE4/4 MCI placebo group at weeks 26, 52, and 78 (all p < 0.001).

Conclusions

Valiltramiprosate 265 mg BID produced early and sustained reductions in plasma p-tau217 and p-tau217/Aβ42 in APOE4 homozygotes and carriers with MCI. In Phase 3 MCI subjects, drug effects on both p-tau217 and NfL correlated significantly with clinical and vMRI benefits, and the two biomarkers were also significantly correlated with each other. These findings suggest that valiltramiprosate engages its central target, thereby reducing Aβ aggregation, tau hyperphosphorylation, and downstream neurodegeneration. The consistent associations of plasma p-tau217 and NfL changes with positive clinical and vMRI outcomes in MCI indicate that the pharmacodynamic biomarker response reflects broader disease-modifying biological effects and are consistent with valiltramiprosate’s mode of action. Together with the previously reported favorable safety and absence of increased ARIA-E risk, these biomarker findings and their clinical correlations support the promising benefit–risk profile of valiltramiprosate in APOE4/4 MCI patients.

Clinical Trial Registry

NCT04770220 and NCT04693520.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s40265-026-02381-3.

Key Points

Amyloid pathology was confirmed in 277 of 294 (94%) participants in the Phase 3 trial and 77 of 80 participants (96%) in the Phase 2 trial with available baseline p-tau217 data.

Valiltramiprosate/ALZ-801 produced early and sustained reductions in plasma p-tau217 and p-tau217/Aβ42 in APOE4/4 homozygotes and APOE4 carriers with mild cognitive impairment (MCI).

In Phase 3 APOE4/4 homozygotes with MCI, drug effects on both p-tau217 and neurofilament light chain levels correlated significantly with cognitive, functional, and brain volumetric MRI benefits, and the two biomarkers were also correlated with each other. In Phase 2 APOE4 carriers, valiltramiprosate produced sustained reductions in plasma p-tau217 over 4 years,  consistent with and supporting the Phase 3 findings.

These biomarker results provide evidence of target engagement and potential disease modification by valiltramiprosate, a novel oral small-molecule inhibitor of amyloid-β oligomer formation in late development as a disease modifying therapy for Alzheimer’s disease (AD), and support its promising clinical efficacy for APOE4/4 homozygotes and APOE4 carriers with MCI due to AD.

Introduction

Alzheimer’s disease (AD), the most common cause of dementia, affects more than 7 million people in the USA and is projected to affect approximately 13 million by 2050 [1]. The apolipoprotein ε4 allele (APOE4) is the strongest genetic risk factor for late-onset AD and has a clear gene–dose effect, including earlier disease onset and faster rate of progression [2, 3]. APOE4/4 homozygotes, who represent approximately 15% of AD cases, show near-complete penetrance of amyloid pathology by age 65 years, higher levels of soluble neurotoxic Aβ oligomers, greater cerebral amyloid angiopathy burden, and accelerated hippocampal atrophy and cortical thinning in early disease compared with APOE3/4 heterozygotes and noncarriers [4, 5]. Patients with APOE4/4 AD are also at greatest risk for amyloid-related imaging abnormalities (ARIA), and show minimal treatment effects with approved amyloid-antibody therapies [6]. This high pathological burden defines APOE4/4 as a population with substantial unmet need for an effective treatment with an improved safety profile, including no increase in ARIA risk.

Within the amyloid/tau/neurodegeneration (A/T/N) framework [7], plasma p-tau217 in the “T” category reflects amyloid-induced tau hyperphosphorylation, a process downstream of the neurotoxic amyloid cascade and upstream of neurofibrillary tangle formation and overt neurodegeneration. Plasma p-tau217 is highly concordant with cerebrospinal fluid (CSF) p-tau and amyloid positron emission tomography (PET), supporting its use as an accessible biomarker for confirming AD pathology [8–12]. In a head-to-head comparison of the major AD blood tests, p-tau217 showed the strongest associations with amyloid PET, tau PET, cortical thickness, and dementia severity, as well as the highest accuracy for classifying amyloid and tau status [13]. The US Food and Drug Administration (FDA) approval of the Fujirebio Lumipulse p-tau217/Aβ42 assay as an aid to AD diagnosis further supports plasma p-tau217 as a practical tool for trial screening and disease monitoring [14].

However, the value of plasma p-tau217 as a pharmacodynamic marker of long-term clinical benefit remains to be fully established across AD treatment mechanisms. In the Phase 2 donanemab trial, substantial plaque clearance and measurable reductions in plasma p-tau217 were not significantly correlated with the primary clinical outcome or whole brain volume changes [15]. We therefore hypothesize that the mechanism driving p-tau217 reduction—plaque clearance versus upstream inhibition of soluble neurotoxic oligomer formation—may determine whether biomarker changes translate into clinical and brain structural benefits.

Neurofilament light chain (NfL), a structural protein of neuronal axons, is released into CSF and plasma after axonal injury and is a marker of neurodegeneration. With sensitive assays such as single-molecule array (Simoa), plasma NfL has become a useful biomarker of neurodegeneration and neuroaxonal injury across neurological disorders. Although not specific to AD, NfL is widely regarded as a reliable plasma biomarker for early detection, monitoring disease progression, and assessing treatment effects in clinical trials.

Valiltramiprosate/ALZ-801 is an oral, brain-penetrant small-molecule inhibitor of soluble neurotoxic Aβ oligomer formation that acts upstream of amyloid plaque deposition and differs mechanistically from plaque-clearing antibodies [16]. It is a valine-conjugated prodrug of tramiprosate that is rapidly converted after oral absorption to tramiprosate and its terminal metabolite, 3-sulfopropanoic acid (3-SPA). Both tramiprosate and 3-SPA inhibit Aβ oligomer formation and together define the primary mechanism of action of valiltramiprosate [17, 18]. By stabilizing Aβ42 monomers in a conformation that prevents oligomer assembly, these agents block synaptic toxicity associated with oligomers and inhibit a proximal driver of amyloid-induced tau hyperphosphorylation. In a Phase 2 trial, valiltramiprosate 265 mg twice daily (BID) produced sustained reductions in plasma p-tau181, reduced hippocampal volume (HV) atrophy, and stabilized cognitive decline over 104 weeks in APOE4 carriers with early AD confirmed by CSF biomarkers [19], while increasing the plasma Aβ42/Aβ40 ratio and slowing the decline in CSF Aβ42 [20]. More recently, we reported topline results from the Phase 3 APOLLOE4 trial, including primary and secondary clinical and imaging outcomes [21]. Although the trial did not meet its primary endpoint in the overall early AD population of APOE4/4 homozygotes, valiltramiprosate produced clinically meaningful, nominally significant benefits in the prespecified MCI subgroup in cognition (ΔADAS-Cog13 = − 2.144, p = 0.041) and function (ΔCDR-SB = − 0.646, p = 0.053) [21]. It also reduced hippocampal atrophy (− 26%, p = 0.0042 in MCI) and showed favorable effects across other brain regions, including greater preservation of cortical thickness and whole brain volume and less ventricular expansion, supporting a neuroprotective effect [21].

In this study, we report plasma biomarker results from the APOLLOE4 Phase 3 trial in APOE4/4 homozygotes with early AD and complementary 4-year fluid biomarker data from a Phase 2 study in APOE4 carriers. We hypothesized that valiltramiprosate’s plasma biomarker effects would depend on disease stage and mirror its clinical and volumetric imaging benefits, with stronger effects in MCI than in Mild AD. Our objectives were to: (1) confirm amyloid AD pathology across both trials using the FDA-approved plasma p-tau217/Aβ42 threshold; (2) assess the effects of valiltramiprosate on plasma p-tau217 and the p-tau217/Aβ42 ratio over 78 weeks in Phase 3 and 208 weeks in Phase 2, and on plasma NfL in Phase 3; and (3) determine whether changes in plasma p-tau217 and NfL correlate with cognitive, functional, and neuroimaging outcomes, with a particular focus on the MCI subgroup, where the clinical, imaging, and biomarker signals were most robust.

Methods

Ethics and Regulatory Compliance

The valiltramiprosate Phase 3 trial was conducted in accordance with the Declaration of Helsinki, International Conference on Harmonization (ICH) guidelines for human research and all applicable national/local regulatory requirements. Central and local institutional review board/independent ethics committees approved the protocol and written informed consent forms. All participants, their legal representative (with participant consent), and trial partners/caregivers provided written informed consent. The Sponsor, Alzheon, designed the trial, provided trial drug, oversaw its conduct, and analyzed the data. An independent Data and Safety Monitoring Board (DSMB) reviewed unblinded safety data every 6 months through closed safety meetings, provided recommendations to the Sponsor, and reviewed the final safety data at the end of trial.

The valiltramiprosate Phase 2 study was conducted at seven sites in the Czech Republic and the Netherlands, after approval of the clinical trial applications by the respective health and regulatory authorities. The study and informed consents were also approved by the Independent Ethics Committee at each study center, and all patients and/or their legal representatives and study partners provided written informed consent. The trial was conducted in accordance with the ICH guidelines for Good Clinical Practice (GCP) and applicable local regulatory requirements.

Valiltramiprosate Phase 3 Trial Design and Participants

The APOLLOE4 valiltramiprosate Phase 3 trial (NCT04770220) was a randomized, double-blind, placebo-controlled study in male and female APOE4/4 homozygotes aged 50–80 years with early AD, comprising both the MCI (MMSE 27–30) and Mild AD (MMSE 22–26) disease stages as determined at the screening visit [21, 22]. This multicenter study was conducted in the USA, Canada, UK, and six EU countries, including the Czech Republic, France, Germany, Iceland, the Netherlands, and Spain. The full enrollment procedures, inclusion and exclusion criteria, and baseline characteristics were previously published [21, 22]. Participants were randomized 1:1 to receive treatment with oral valiltramiprosate tablets 265 mg BID (n = 163) or matching placebo (n = 162) for 78 weeks, with an initial 2-week titration from once daily (QD) to BID dosing. The primary clinical efficacy outcome was the 13-item Alzheimer’s Disease Assessment Scale-Cognitive Subscale (ADAS-Cog13). The secondary outcomes included the Clinical Dementia Rating-Sum of Boxes (CDR-SB), a cognitive/functional composite, and the functional Disability Assessment for Dementia (DAD, previously used in the tramiprosate clinical trials). Longitudinal brain volumetric magnetic resonance imaging (vMRI) assessments were also conducted. Bilateral HV was the prespecified main imaging outcome, with cortical thickness as secondary, and ventricular volume and whole brain volume as additional imaging outcomes. The topline clinical, neuroimaging, pharmacokinetic (PK), and safety results were recently reported [21].

Valiltramiprosate Phase 2 Trial Design and Participants

The single-arm, open-label Phase 2 biomarker study (NCT04693520) was designed to evaluate the longitudinal effects of valiltramiprosate 265 mg BID on biomarkers of AD pathology in APOE4 carriers with early AD, including 31 APOE4/4 homozygotes and 53 APOE3/4 heterozygotes. The study was conducted at seven study sites, three in the Netherlands and four in the Czech Republic. All participants had confirmed positive biomarkers of amyloid and tau pathologies in the CSF (i.e., high p-tau181 and low Aβ42/Aβ40 levels), ensuring a highly amyloid-enriched population at baseline. The full enrollment procedures, inclusion and exclusion criteria, and baseline characteristics have been previously published [19]. Following a 2-week QD titration during the first 2 weeks, all participants received valiltramiprosate tablets 265 mg BID for 104 weeks during the main study. Subsequently, most of the subjects continued with the same treatment regimen for another 2 years during the long-term extension phase. The primary efficacy outcome was the change from baseline in plasma p-tau181 over 104 weeks; HV evaluated by vMRI was the key secondary outcome, and plasma Aβ42 and Aβ40 were the secondary biomarker outcomes. The clinical endpoints included the Rey Auditory Verbal Learning Test (RAVLT), CDR-SB and Digit Symbol Substitution Test (DSST). A composite cognitive score combining the RAVLT total memory scores (immediate recall, delayed recall, and recognition) and the DSST score (named 3-item RAVLT+DSST) was calculated to bridge to the Phase 3 cognitive data with ADAS-Cog13. The topline clinical, biomarker (plasma p-tau181, CSF Aβ42, and plasma Aβ42/Aβ40 ratio), neuroimaging, PK and safety results from the main study (i.e., 104 weeks) were previously reported [19, 20, 23]. Plasma p-tau217 (by Fujirebio Lumipulse assay), which was not available at the time of study completion and thus was not initially measured, was assayed at the same time with the Phase 3 plasma samples in late 2025 and reported in the present analysis.

Plasma p-Tau217, p-Tau217/Aβ42 and NfL Measurements

Levels of p-tau217 and p-tau217/Aβ42 ratio were quantified in K3EDTA plasma samples from all subjects with available longitudinal plasma samples in the valiltramiprosate Phase 3 and Phase 2 trials, using the FDA-approved Fujirebio Lumipulse G assays at Frontage Laboratories (Exton, PA, USA). The bioanalysis was conducted in a blinded manner, in compliance with the standard operating procedures at Frontage Laboratories and applicable regulatory guidelines. The assays are based on the chemiluminescent enzyme immunoassay (CLEIA) technology using a specific two-step immunoassay method using the Lumipulse G System. Briefly, the procedure begins with adding p-tau217 or Aβ42 clinical specimens or plasma calibrators into a particle solution, where p-tau217 or Aβ42 specifically binds to its monoclonal antibody (mouse) on the particles, followed by the formation of antigen–antibody immunocomplexes. The particles are then washed and rinsed to remove the unbound materials. An alkaline phosphatase (ALP)-labeled anti-p-tau217 or Aβ42 monoclonal antibody (mouse) is then added which specifically binds to the prior formed immunocomplexes on the particles, and additional immunocomplexes are formed. The particles are again washed and rinsed to remove the unbound materials. Subsequently, a substrate solution containing AMPPD (3-(2’-spiroadamantane)-4-methoxy-4-(3”-phosphoryloxy)-phenyl-1, 2-dioxetane disodium salt) is added, which is dephosphorylated by the catalysis of ALP indirectly conjugated to the particles. Luminescence (at a maximum wavelength of 477 nm) is generated by the cleavage reaction of dephosphorylated AMPPD, which is then read by the Lumipulse G1200 instrument. The raw data are then fitted with a built-in linear regression model using the Fujirebio software, with the analyte concentrations reported and p-tau217/Aβ42 ratios calculated. AD amyloid positivity was defined using the FDA diagnostic plasma p-tau217/Aβ42 ratio cut-off ≥ 0.00738 [25].

Both the p-tau217 and Aβ42 methods were pre-validated at Frontage Laboratories to the manufacturer’s specifications immediately prior to the assay of the present clinical plasma samples. For p-tau217, the standard curve calibration range was 0.000–10.000 pg/mL. The intra-assay coefficient of variation (CV) was 0.3–3.7%, and the inter-assay CV was 1.8–3.8%. For Aβ42, the standard curve calibration range was 0.000–1000.0 pg/mL. The intra-assay CV was 0.2–3.0%, and the inter-assay CV was 2.4–4.1%. Both assays were confirmed without apparent matrix interferences (1–2% hemolysis and 300 mg/dL triglycerides) or freeze-thaw effects up to five cycles.

Plasma NfL levels in the valiltramiprosate Phase 3 trial were quantified using the high-sensitivity Simoa immunoassay at Quanterix Laboratories (Billerica, MA, USA). In brief, this technology makes use of arrays of femtoliter-sized reaction chambers, termed single-molecule arrays, that have the capability to isolate and detect fluorescence from a single labeled protein molecule. Human plasma samples were assayed in duplicate, with the concentrations determined using a four-parameter logistic calibration curve fitted as per the manufacturer’s specifications. The LLOQ was 0.400 pg/mL and the ULQQ was 558 pg/mL. The intra-assay CV was 4.5–5.3% and the inter-assay CV was 5.2–6.4%.

Volumetric MRI Measures

The vMRI methods of the ALZ-801 Phase 2 and 3 studies have been previously reported [19, 21]. Bilateral vMRI measures and change from baseline (CBL) analyses were conducted in a blinded manner by Clario (Philadelphia, PA, USA). For correlation analyses with the plasma biomarkers, the vMRI change from baseline utilized boundary shift integral (BSI) method [19, 21, 24]. Both the baseline and CBL data were analyzed; CBL was defined as the observed value at a given study visit minus the baseline value.

Data Analyses

Plasma biomarker analyses were prespecified outcomes in the Phase 3 trial [21] and were the main objective and outcomes of the Phase 2 biomarker trial [19]. In this a priori analysis, observed case analyses were conducted for the plasma biomarker endpoints for the overall population, and the prespecified MCI and Mild AD subgroups. Subject-level correlations were assessed using Spearman’s correlation coefficients between baseline or CBL in plasma biomarkers and baseline or CBL in clinical and vMRI imaging outcomes. In AD studies, biomarker-clinical or biomarker-vMRI correlations between 0.2 and 0.29 are generally considered modest, between 0.3 and 0.39 are considered moderate, between 0.4 and 0.49 are considered strong, and ≥ 0.5 are considered very strong [15].

Between-group p-values for observed case plasma biomarkers at each visit were assessed using the unpaired, two-sided t-test. Where appropriate, multiple group comparisons were analyzed using one-way analysis of variance (ANOVA) followed by Dunnett’s multiple comparison test. Statistical significance was defined as p < 0.05.

Results

We report plasma p-tau217, p-tau217/Aβ42 and NfL results from a total of 294 APOE4/4 early AD subjects (n = 144 for valiltramiprosate 265 mg BID treatment and n = 150 for placebo treatment) in the Phase 3 APOLLOE4 trial with available longitudinal plasma samples for the biomarker assays. Additionally, plasma p-tau217 and p-tau217/Aβ42 results from 80 APOE4 carrier subjects with early AD (all with valiltramiprosate 265 mg BID treatment) in the Phase 2 biomarker trial (core study + 2-year long-term extension) are also reported; 49 subjects completed 4 years of treatment with 31 early terminations. Tables 1 and 2 summarize the demographic and baseline clinical characteristics of the subjects from the Phase 3 and Phase 2 studies that were included in the plasma biomarker analysis.

Table 1.

Demographic and baseline characteristics of the Phase 3 study set of subjects in the plasma biomarker analysis

Characteristic	Valiltramiprosate (n = 144)	Placebo (n = 150)
Age, mean ± SD (years)	68.4 ± 6.3	68.8 ± 6.0
Sex, female, n (%)	73 (50.7%)	73 (48.7%)
Race, white, n (%)	126 (87.5%)	137 (91.3%)
Ethnicity, non-Hispanic/Latino, n (%)	129 (89.6%)	132 (88.0%)
Body mass index, mean ± SD (kg/m2)	26.0 ± 4.0	25.0 ± 4.4
Diagnosis, MCI, n (%)	60 (41.7%)	56 (37.3%)
Acetylcholinesterase inhibitor use	49 (34.0%)	58 (38.7%)
Screening MMSE (visit 1a), mean ± SD	25.8 ± 2.6	25.6 ± 2.4
Randomization MMSE (visit 2), mean ± SD	25.4 ± 3.1	24.8 ± 3.5
ADAS-Cog13, mean ± SD	23.4 ± 8.3	24.1 ± 8.5
CDR-G, mean ± SD	0.6 ± 0.2	0.6 ± 0.2
CDR-SB, mean ± SD	3.0 ± 1.5	3.0 ± 1.5
DAD, mean ± SD	36.5 ± 4.2	36.3 ± 4.4

The screening period included two visits (visit 1a and visit 2) to ensure participant eligibility, with a maximum of 13 weeks allowed for the screening period (between visit 1a and visit 2)

ADAS-Cog13 Alzheimer’s Disease Assessment Scale-cognitive subscale 13 item, with the total score ranges from 0 to 85 and higher scores indicate greater impairment, CDR-G Clinical Dementia Rating–Global Score, CDR-SB Clinical Dementia Rating–Sum of Boxes with the total score ranges from 0 to 18 and higher scores indicate greater impairment, DAD Disability Assessment for Dementia with the total score ranges from 0 to 100 and higher score indicates greater function competence, MCI mild cognitive impairment, defined as an MMSE score of 27–30, MMSE Mini-Mental State Examination, which ranges from 0 to 30 with higher scores indicating better cognition

Table 2.

Demographic and baseline characteristics of the Phase 2 study set of subjects in the plasma biomarker analysis

Characteristics	Valiltramiprosate (n = 80)
Age, mean ± SD (years)	68.8 ± 5.4
Sex, female, n (%)	41 (51.3%)
Race, white, n (%)	80 (100%)
Body mass index, mean ± SD (kg/m2)	25.2 ± 3.2
APOE 3/4, n (%)	50 (62.5%)
APOE 4/4, n (%)	30 (37.5%)
MCI, n (%)	35 (43.7%)
Mild AD, n (%)	45 (56.3%)
Acetylcholinesterase inhibitor use	45 (56.3%)
MMSE, mean ± SD	25.8 ± 2.6
CDR-G, mean ± SD	0.6 ± 0.2
CDR-SB, mean ± SD	2.7 ± 1.4
RAVLT-Total, mean ± SD	30.5 ± 13.1
DSST, mean ± SD	38.9 ± 18.2

MMSE Mini-Mental State Examination, MCI mild cognitive impairment, CDR-G Clinical Dementia Rating Scale-Global Score, CDR-SB Clinical Dementia Rating Scale-Sum of Boxes, DSST Digit Symbol Substitution Test, RAVLT Rey Auditory Verbal Learning Test, RAVLT-Total score RAVLT-immediate memory score + RAVLT-delayed memory score

Confirmation of AD Pathology in Valiltramiprosate Phase 3 and Phase 2 Trials

AD amyloid pathology was confirmed in the majority of the participants enrolled across both trials using the FDA-approved Fujirebio Lumipulse plasma p-tau217/Aβ42 assay cut-off threshold of 0.00738 [25]. In the Phase 3 APOLLOE4 trial, 94% (n = 277 out of 294) of the enrolled APOE4/4 subjects met the amyloid positivity threshold across the overall early AD population, and 91% (n = 105 out of 116) of the MCI subjects were confirmed to be amyloid positive. In the Phase 2 biomarker trial, which enrolled subjects based on the CSF biomarker inclusion criteria (i.e., high p-tau181 and low Aβ42/Aβ40 ratio [19]) and therefore represented a highly amyloid-enriched population at study entry, 97% of APOE4/4 homozygotes (n = 29 out of 30) and 96% of APOE3/4 heterozygotes (n = 48 out of 50) met the plasma AD positivity threshold. Therefore, the high concordance between the plasma p-tau217/Aβ42 threshold and the CSF-based inclusion criteria in Phase 2 provides retrospective validation of the new plasma assay sensitivity against a gold-standard enriched reference population.

Plasma p-Tau217, p-Tau217/Aβ42, and Aβ42 Changes Over 78 Weeks in Valiltramiprosate Phase 3 Overall, MCI, and Mild AD Populations

In the Phase 3 overall APOE4/4 population with early AD, baseline plasma p-tau217 concentrations were comparable between the treatment groups (0.644 ± 0.099 pg/mL for valiltramiprosate, n = 144 and 0.623 ± 0.032 pg/mL for placebo, n = 150; mean ± SEM, p = 0.837). Beginning at week 26 (i.e., first study visit with analyzed plasma biomarkers), the two study arms exhibited divergent trajectories: valiltramiprosate-treated participants showed a sustained reduction from baseline while placebo-treated participants showed a progressive elevation (Fig. 1). The between-group difference reached significance at week 26 (p = 0.021), week 52 (p = 0.014), and week 78 (p = 0.011), with the valiltramiprosate-treated participants showing observed values of 0.606 ± 0.039 pg/mL and placebo-treated participants reaching 0.755 ± 0.043 pg/mL by week 78.

Fig. 1.

Longitudinal changes in plasma p-tau217 concentrations (pg/mL) over time following valiltramiprosate (ALZ-801) 265 mg BID versus placebo treatment in the overall APOE4/4 population of the Phase 3 trial. Data are shown as observed case analysis; means ± SEM, n = 144 for valiltramiprosate and n = 150 for placebo. *p < 0.05, ALZ-801 versus placebo, unpaired, two tailed t-test. These data show a rapid and sustained reduction of plasma p-tau217 levels after valiltramiprosate treatment versus a progressive increase with placebo over 78 weeks in the overall APOE4/4 population with early Alzheimer’s disease

In the MCI subgroup, baseline plasma p-tau217 concentrations were numerically higher in the valiltramiprosate arm than in placebo (0.689 ± 0.230 pg/mL, n = 60 versus 0.524 ± 0.051 pg/mL, n = 56, respectively), although this difference was not statistically significant (p = 0.499). Following the initiation of treatment, plasma p-tau217 in valiltramiprosate-treated MCI subjects declined sharply from baseline, while placebo-treated MCI subjects showed a progressive increase over the same time interval (Fig. 2). Of note, a directional separation between the two treatment arms was apparent as early as week 26 (p = 0.097), with the between-group difference reaching significance at week 52 (p = 0.030) and week 78 (p = 0.020). By week 78, valiltramiprosate-treated MCI subjects showed a 36% decrease in plasma p-tau217 levels from baseline, whereas placebo-treated MCI subjects had a 17% increase, corresponding to an approximately 53% difference versus placebo. In Mild AD, valiltramiprosate-treated subjects showed numerically lower plasma p-tau217 levels relative to placebo across all timepoints; however, the between-group difference did not reach statistical significance (Supplementary Data Fig. S1).

Fig. 2.

Longitudinal changes in plasma p-tau217 concentrations (pg/mL) over time following valiltramiprosate (ALZ-801) 265 mg BID versus placebo treatment in the mild cognitive impairment (MCI) APOE4/4 population of the Phase 3 trial. Data are shown as observed case analysis; means ± SEM, n = 60 for valiltramiprosate and n = 56 for placebo. *p < 0.05, ALZ-801 versus placebo, unpaired, two tailed t-test. These data show a rapid and sustained reduction of plasma p-tau217 levels after valiltramiprosate treatment versus a progressive increase with placebo over 78 weeks in the APOE4/4 population with MCI

Consistent with plasma p-tau217, the plasma p-tau217/Aβ42 ratio showed a parallel pattern in the overall, MCI and Mild AD subjects (Supplementary Data Fig. S2). Over 78 weeks of study duration, valiltramiprosate-treated MCI subjects showed a 40% decrease in the p-tau217/Aβ42 ratio from baseline, while placebo-treated MCI subjects showed a 7% increase (Supplementary Data Fig. S3).

Longitudinal plasma Aβ42 results in the overall, MCI and Mild AD populations are summarized in Supplementary Data Table S1. No significant changes were observed for this biomarker alone.

Correlations between Plasma p-Tau217 and Clinical and Neuroimaging Outcomes in Valiltramiprosate Phase 3 MCI Subjects

To understand the relationship between the plasma biomarker effects and the clinical and neuroimaging outcomes following valiltramiprosate treatment, we conducted Spearman’s correlation analyses between the changes from baseline (CBL) in plasma p-tau217 levels and CBL in clinical and brain vMRI outcomes at week 78 in the Phase 3 study. In the overall population (Table 3), plasma p-tau217 reduction at week 78 was significantly associated with the improvements in cognitive/functional and vMRI outcome measures (CDR-SB, HV, cortical thickness, and ventricular volume; all p < 0.05) after valiltramiprosate treatment. There was a trend effect in the association with ADAS-Cog13 and whole brain volume (p < 0.10). In contrast, there were fewer significant correlations in the overall placebo group (CDR-SB and ventricular volume; both p < 0.05). In the MCI population (Table 4), plasma p-tau217 reduction at week 78 was significantly correlated with the improvements in cognitive/functional and vMRI outcome measures (ADAS-Cog13, CDR-SB, HV, cortical thickness, and whole brain volume; all p < 0.05) after valiltramiprosate treatment. Also, there was a trend effect in the association with ventricular volume (p = 0.074). In contrast, no correlation was found with the clinical outcomes or vMRI outcome measures in the placebo MCI subjects (all p > 0.05). In Mild AD (Supplementary Data Table S2), no significant correlation was found in the valiltramiprosate-treated subjects (all p > 0.05). There were two significant correlations (CDR-SB and ventricular volume; both p < 0.05) in the placebo Mild AD subjects. Taken together, these correlation analyses showed a consistent, multidomain pharmacodynamic signal linking plasma p-tau217 reduction to the preservation of cognition, function and brain structures in both the overall and MCI APOE4/4 subjects after valiltramiprosate treatment.

Table 3.

Correlations between the plasma p-tau217 drug effect (change from baseline at week 78) and the clinical and neuroimaging outcomes (change from baseline at week 78) in Phase 3 APOE4/4 subjects in the overall population: valiltramiprosate treatment versus placebo

	Valiltramiprosate			Placebo
Outcome	n	r	p-value	n	r	p-value
Positive correlation coefficient r values consistent with association
ADAS-Cog 13	128	0.148	0.096	138	0.117	0.170
CDR-SB	124	0.216	0.016	136	0.270	0.002
Ventricular volume	109	0.191	0.047	119	0.211	0.021
Negative correlation coefficient r values consistent with association
DAD	127	− 0.121	0.174	138	− 0.113	0.188
Hippocampal volume	109	− 0.196	0.041	119	− 0.139	0.131
Cortical thickness	109	− 0.232	0.015	119	− 0.114	0.218
Whole brain volume	109	− 0.16	0.096	119	− 0.148	0.108

Significant correlations with p < 0.05 are shown in bold

For plasma p-tau217, ADAS-Cog13, CDR-SB, and ventricular volume, larger values indicate more severe disease. For DAD, hippocampal volume, cortical thickness and whole brain volume, smaller values indicate more severe disease. These results show more significant correlations of plasma p-tau217 changes and clinical/neuroimaging outcomes at week 78 in valiltramiprosate-treated overall subjects than in placebo subjects, especially in hippocampal volume and cortical thickness

ADAS-Cog13 13-item Alzheimer’s Disease Assessment Scale-Cognitive Subscale, CDR-SB Clinical Dementia Rating-Sum of Boxes, DAD Disability Assessment for Dementia, r Spearman’s correlation coefficient

Table 4.

Correlations between the plasma p-tau217 drug effect (change from baseline at week 78) and the clinical and neuroimaging outcomes (change from baseline at week 78) in Phase 3 APOE4/4 subjects with MCI: valiltramiprosate treatment versus. placebo

	Valiltramiprosate			Placebo
Outcome	n	r	p-value	n	r	p-value
Positive correlation coefficient r values consistent with association
ADAS-Cog 13	56	0.276	0.039	52	0.098	0.490
CDR-SB	54	0.377	0.005	51	0.198	0.164
Ventricular volume	49	0.257	0.074	44	0.164	0.287
Negative correlation coefficient r values consistent with association
DAD	56	− 0.117	0.391	52	− 0.062	0.660
Hippocampal volume	49	− 0.353	0.013	44	− 0.032	0.838
Cortical thickness	49	− 0.337	0.018	44	− 0.117	0.449
Whole brain volume	49	− 0.312	0.029	44	− 0.080	0.606

Significant correlations with p < 0.05 are shown in bold

For plasma p-tau217, ADAS-Cog13, CDR-SB and ventricular volume, larger values indicate more severe disease. For DAD, hippocampal volume, cortical thickness and whole brain volume, smaller values indicate more severe disease. These results show multiple significant correlations of plasma p-tau217 changes and clinical/neuroimaging outcomes at week 78 in valiltramiprosate-treated MCI subjects, but not in placebo MCI subjects

ADAS-Cog13 13-item Alzheimer’s Disease Assessment Scale-Cognitive Subscale, CDR-SB Clinical Dementia Rating-Sum of Boxes, DAD Disability Assessment for Dementia, r Spearman’s correlation coefficient

To further investigate the physiological relationship of plasma p-tau217 and the clinical and neuroimaging measures, we conducted cross-sectional associations between baseline plasma p-tau217 levels and baseline clinical and brain vMRI measures in the overall MCI population (i.e., combined valiltramiprosate and placebo subjects at baseline). As presented in Supplementary Data Table S3, significant correlations between plasma p-tau217 levels and ADAS-Cog13, CDR-SB, DAD, HV, cortical thickness, and ventricular volume were observed (all p < 0.05). While these predose baseline results do not suggest drug effects, they nonetheless indicate an important role of plasma p-tau217 levels as a pathophysiological biomarker of disease stage in AD progression.

Plasma NfL Changes and Correlations with Clinical and Neuroimaging Outcomes in Valiltramiprosate Phase 3 Overall, MCI, and Mild AD Subjects

We analyzed the changes of plasma NfL, as measured by the Simoa assay, in the Phase 3 study. As shown in Fig. 3, in the overall population, there was a trend to lower plasma NfL levels with valiltramiprosate than placebo at both baseline and week 78, although the difference did not reach statistical significance. In the MCI subgroup (Fig. 4), baseline NfL values were not significantly different between the treatment arms (16.7 ± 1.0 pg/ml and 15.0 ± 0.7 pg/ml for placebo and valiltramiprosate, respectively; mean ± SEM). At week 78, however, the valiltramiprosate arm showed significantly lower NfL levels than placebo (18.9 ± 1.0 pg/ml and 16.1 ± 0.8 pg/ml for placebo and valiltramiprosate, respectively, mean ± SEM; p = 0.032). In Mild AD subjects, no significant difference on plasma NfL levels was observed between valiltramiprosate and placebo (Supplementary Data Fig. S4).

Fig. 3.

Plasma neurofilament light chain (NfL) concentrations (pg/mL) at week 78 following valiltramiprosate (ALZ-801) 265 mg BID versus placebo treatment in the overall APOE4/4 population of the Phase 3 trial. Data are shown as observed case analysis; means ± SEM, n = 154 for valiltramiprosate and n = 151 for placebo at baseline. Valiltramiprosate treatment showed a nonsignificant, numerical trend of lower plasma NfL versus placebo at 78 weeks

Fig. 4.

Plasma neurofilament light chain (NfL) concentrations (pg/mL) at week 78 following valiltramiprosate (ALZ-801) 265 mg BID versus placebo treatment in the mild cognitive impairment (MCI) APOE4/4 population of the Phase 3 trial. Data are shown as observed case analysis; means ± SEM, n = 65 for valiltramiprosate and n = 56 for placebo. *p < 0.05, ALZ-801 versus placebo, unpaired, two tailed t-test. Left panel, bar graph showing the concentrations at both baseline and week 78. Right panel, individual plot showing the concentration results at week 78. One outlier value was excluded from the analysis due to value ≥ 3× standard deviation. Valiltramiprosate treatment showed significantly lower plasma NfL versus placebo at 78 weeks in APOE4/4 subjects with MCI (p = 0.032)

As NfL is a direct marker of neuroaxonal injury and degeneration, we conducted correlation analyses of plasma NfL versus brain vMRI measures in MCI subjects. As shown in Supplementary Data Fig. S5, there were significant correlations between plasma NfL CBL and HV CBL at both 52 weeks (r = − 0.455, p = 0.0008) and 78 weeks (r = − 0.282, p = 0.047) following valiltramiprosate treatment. In contrast, no such effect was observed in the placebo MCI subjects. Similarly, significant correlations were also seen when plasma NfL CBL was analyzed with cortical thickness and whole brain volume CBL in MCI subjects with valiltramiprosate treatment (data on file, Alzheon Inc.). In addition, valiltramiprosate drug effects on plasma NfL CBL correlated significantly with drug effects on ADAS-Cog13 CBL (r = 0.29, p = 0.03) and on p-tau217 CBL (r = 0.42, p = 0.001) in MCI subjects, indicating that individuals with greater p-tau217 reductions also showed greater NfL stabilization, reflecting convergent effects on tau phosphorylation and neuroaxonal loss (neurodegeneration).

Valiltramiprosate Phase 2 Biomarker Study: Plasma p-Tau217 Changes and Long-Term Clinical and Neuroimaging Outcomes

Figure 5 shows the Phase 2 plasma p-tau217 results in APOE4 carriers assayed at the same time as the Phase 3 study. There was a significant reduction in plasma p-tau217 that occurred as early as 26 weeks, the first sampling visit after valiltramiprosate dosing. The reduction (approximately 45%) from baseline was the largest at week 52, which was sustained below baseline through 208 weeks (i.e., 4 years including the long-term extension). When compared by disease severity, both the MCI and Mild AD subjects showed plasma p-tau217 reduction, with a greater effect in the former (Supplementary Data Fig. S6). The longitudinal temporal profile of p-tau217 reduction in the Phase 2 MCI groups was similar to that observed in the Phase 3 trial above.

Fig. 5.

Longitudinal changes in plasma p-tau217 concentrations (pg/mL) over 208 weeks following valiltramiprosate (ALZ-801) 265 mg BID treatment in overall APOE4 carriers with early Alzheimer’s disease (AD) in the single-arm Phase 2 trial. Data are shown as observed case analysis; means ± SEM, n = 80, all received valiltramiprosate treatment. *p < 0.05, **p < 0.01, and ***p < 0.001 versus baseline, one-way ANOVA with Dunnett’s multiple comparison post-hoc test. These results show a rapid reduction of plasma p-tau217 levels after valiltramiprosate treatment which was sustained and remained below baseline over the 4-year study duration in APOE4 carriers with early AD

While this Phase 2 biomarker trial was conducted as a single-arm study without a placebo arm (i.e., all study participants received valiltramiprosate 265 mg BID), we compared the Phase 2 plasma p-tau217 results in APOE4 carriers with MCI to the placebo data from the Phase 3 MCI APOE4/4 subjects (Fig. 6). Compared with the Phase 3 placebo MCI arm, plasma p-tau217 levels were significantly lower with valiltramiprosate treatment in both APOE4 carriers and heterozygotes with MCI (p < 0.001 at all time points at weeks 26, 52, and 78).

Fig. 6.

Longitudinal changes in plasma p-tau217 concentrations (pg/mL) over time following valiltramiprosate (ALZ-801) 265 mg BID in APOE4 carriers and heterozygotes with mild cognitive impairment (MCI) in the Phase 2 trial, compared with the placebo MCI arm from the Phase 3 trial. Data are shown as observed case analysis; means ± SEM. **p < 0.01, ALZ-801 versus placebo, unpaired, two tailed t-test. A Comparison of plasma p-tau217 changes in the overall APOE4 carriers with MCI (n = 80) in the Phase 2 trial versus the placebo MCI arm (n = 56) from the Phase 3 trial over 78 weeks. B Comparison of plasma p-tau217 changes in APOE4 heterozygotes with MCI (n = 50) in the Phase 2 trial versus the placebo MCI arm (n = 56) from the Phase 3 trial over 78 weeks. These results show significantly lower plasma p-tau217 levels at each timepoint after valiltramiprosate treatment in both Phase 2 APOE4 carriers and heterozygotes with MCI when compared with an external placebo MCI APOE4/4 group from the Phase 3 trial

Consistent with the plasma p-tau217 findings, MCI APOE4 carriers in the Phase 2 biomarker trial showed cognitive stability up to 4 years on the composite RAVLT+DSST assessment (Fig. 7). In addition, Table 5 presents the correlations of plasma p-tau217 CBL at week 208 with clinical and HV outcomes CBL at week 208 in the Phase 2 overall MCI APOE4 carrier population. A greater reduction in plasma p-tau217 in MCI subjects correlated significantly with better RAVLT-Total score (r = − 0.537, p = 0.006), composite RAVLT+DSST score (r = − 0.477, p = 0.021), and CDR-SB (r = 0.464, p = 0.019) at week 208. However, the correlation with HV changes at week 208 was not significant, suggesting that HV trajectories over very long follow-up periods might diverge from the plasma p-tau217 trajectory, which stabilizes after the initial reduction at week 26.

Fig. 7.

Longitudinal changes in the composite cognitive score 3-item RAVLT+DSST over 4 years following valiltramiprosate (ALZ-801) 265 mg BID treatment in APOE4 carriers with mild cognitive impairment (MCI) in the Phase 2 trial. Data are shown as observed case analysis; means ± SEM. These results show cognitive stabilization in APOE4 carriers with MCI over 4 years after valiltramiprosate treatment

Table 5.

Correlations between the plasma p-tau217 drug effect (change from baseline at week 208) and the clinical and hippocampal volume outcomes (change from baseline at week 208) in Phase 2 APOE4 carriers with MCI receiving valiltramiprosate treatment

	Valiltramiprosate
Outcome	n	r	p-value
Positive correlation coefficient r values consistent with association
CDR-SB	25	0.464	0.019
Negative correlation coefficient r values consistent with association
RAVLT-Total	25	− 0.537	0.006
3-Item RAVLT+DSST	23	− 0.477	0.021
Hippocampal volume	14	− 0.182	0.533

Significant correlations with p < 0.05 are shown in bold

For plasma p-tau217 and CDR-SB, larger values indicate more severe disease. For RAVLT-Total, 3-Item RAVLT+DSST, and hippocampal volume, smaller values indicate more severe disease. All participants in the Phase 2 study received valiltramiprosate 265 mg BID treatment. These results show significant correlations of plasma p-tau217 changes and cognitive/functional outcomes following 4 years of valiltramiprosate treatment in APOE4 carriers with MCI

CDR-SB Clinical Dementia Rating-Sum of Boxes, r Spearman’s correlation coefficient, RAVLT-Total Rey Auditory Verbal Learning Test Total score (including immediate recall, delayed recall and recognition), 3-Item RAVLT+DSST a composite cognitive score combining the RAVLT-Total score and the Digit Symbol Substitution Test (DSST) score

Discussion

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.

Conclusions

Valiltramiprosate 265 mg BID produced early and sustained reductions in plasma p-tau217 in APOE4/4 homozygotes and APOE4 carriers with MCI, and these reductions correlated significantly with cognitive, functional, and neuroimaging outcomes. These findings are consistent with pharmacodynamic target engagement of the amyloid pathway through upstream inhibition of neurotoxic soluble amyloid oligomer formation, leading to downstream effects on tau phosphorylation, neurodegeneration (as reflected by NfL and brain regional atrophy), and slowing clinical decline. The high rates of amyloid AD pathology confirmation across both trials further support the enrichment strategies used. The previously reported favorable safety profile and absence of ARIA-E risk are consistent with valiltramiprosate’s upstream mechanism of action and its lack of interaction with, or active removal of, insoluble amyloid plaques. Together, these biomarker findings and their significant correlation to clinical benefit support the promising profile of valiltramiprosate and its planned evaluation in biomarker-enriched Phase 3 trials in APOE4/4 homozygotes and APOE4 carriers with MCI or early symptomatic AD.

Supplementary Information

Below is the link to the electronic supplementary material.

Funding

The APOLLOE4 Phase 3 trial was funded by a grant award no. R01AG065253 from the NIH/NIA and by the sponsor, Alzheon, Inc. The Phase 2 biomarker trial was funded by Alzheon, Inc. Publication of this article was funded by Alzheon, Inc. The content is solely the responsibility of the authors and does not necessarily represent the official views of the funding agency.

Declarations

Ethics Approval

The APOLLOE4 trial was approved by Advarra Central IRB and by the required local university IRBs in the USA, and by country specific Ethics Committees in UK and EU. The study was performed in accordance with the ethical standards as laid down in the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards. The Phase 2 study protocol was approved by the ethics committees (EC) in the Netherlands and in Czech Republic. The study was conducted in accordance with the International Council for Harmonization (ICH) Guideline for Good Clinical Practice (GCP) and applicable local regulatory requirements.

Consent to Participate

All Phase 3 study subjects and their partners, caregivers, or legal representatives signed informed consent before participation in the trial. The Phase 2 study and informed consents were also approved by the Independent Ethics Committee at each study center, and all patients and/or their legal representatives and study partners provided written informed consent.

Consent for Publication

Not applicable.

Code Availability

Not applicable.

Data Availability

The data supporting findings of this study are available from Alzheon, Inc., but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are, however, available from the corresponding authors upon reasonable request and with permission of Alzheon, Inc.

Conflict of interest

JA Hey, S Abushakra, A Power, E Ling, S Flint, E Hristova, P Kesslak, R McLaine, A Albayrak, JF Schaefer, JY Yu, L Tolar, and M Tolar are employees of Alzheon Inc., who received salary compensations, stocks and/or stock options from Alzheon Inc. The AD experts (A Porsteinsson, M Sabbagh, PM Doraiswamy, FM Elahi and RA Rissman) received consulting fees, stocks, and/or stock options from Alzheon Inc. and multiple other pharmaceutical companies developing CNS drugs, devices or diagnostic tests. The site investigators (D Watson, E MacSweeney, S Cohen and M Boada) received investigator fees for this study and multiple other AD and CNS trials from various global pharmaceutical companies.

Author Contributions

JA Hey, JY Yu, and L Tolar contributed to manuscript preparation. S Abushakra is the principal investigator of the NIH/NIA grant for the APOLLOE4 trial. S Abushakra, A Power, E Liang, S Flint, JA Hey, and M Tolar designed the trial and oversaw its conduct, data analysis, data interpretation, and manuscript preparation. All other Alzheon employees contributed to one or more of the following: trial conduct and oversight, safety oversight, sample analyses, data analyses, and data interpretation. A Porsteinsson, M Sabbagh, PM Doraiswamy, FM Elahi, and RA Rissman were scientific advisors to Alzheon Inc. who contributed to the study design, data analysis, data interpretation, and critical review of the manuscript. D Watson, E McSweeney, S Cohen, and M Boada were trial investigators in the USA, UK, Canada, and Europe, respectively; they contributed to the trial conduct, safety discussions, data interpretation and manuscript review. All authors reviewed and approved the final submitted version of the manuscript.