Introduction
Alzheimer’s disease (AD) is a major health problem of aging with tremendous burden on healthcare systems globally [
1‐
4]. Although there are symptomatic therapies approved, they provide modest clinical benefit and have no impact on disease progression. One anti-amyloid antibody (aducanumab) has accelerated approval from the FDA; it does not fulfill the unmet need for AD therapy. Therefore, disease-modifying therapies are critically needed to improve the lives of those with AD and to decrease the global burden of the disease.
Amyloid beta (Aβ) pathology has been identified as a target for intervention based on the evidence that it likely plays an important role in the development and progression of the disease. However, nearly all symptomatic AD trials targeting Aβ-pathology have been unsuccessful in demonstrating a clinical benefit [
5‐
7]. Only recently have specific therapies that effectively target Aβ-pathology been developed and include those that substantially reduce aggregated Aβ plaques, decrease soluble Aβ, or reduce the production of aggregation prone Aβ species [
8‐
11]. Although it remains uncertain which form of aggregated Aβ is likely to be most pathologic, soluble aggregates (e.g., large, soluble protofibrils) are a rational target based on evidence that these may be the most toxic forms [
12‐
14]. Lecanemab (BAN2401) is a humanized IgG1 monoclonal antibody that binds to large soluble Aβ aggregates (protofibrils) with high selectivity over monomers (>1000-fold) and insoluble fibrils (at least 10-15-fold) [
15‐
18].
In a phase 2 randomized study (study 201 core), lecanemab treatment led to a robust, dose-dependent reduction in brain amyloid, slower decline on clinical outcome measures, and directionally consistent biomarker changes at 18 months [
19]. An open-label extension (OLE) was initiated following analysis of core data, resulting in a gap period (no study drug treatment) between the end of the core and the beginning of the OLE. Here, we report detailed results from study 201 core, gap period, and OLE phase supporting the effectiveness of lecanemab, including plasma biomarker outcomes, clinical efficacy, and exposure response (ER) data as well as correlations among positron emission tomography (PET) measures, plasma biomarker assessments, and clinical efficacy evaluations. This study addressed efficacy only and does not present data on safety or tolerability which are addressed elsewhere [
19].
Patients and methods
Study design
The lecanemab 201 trial (ClinicalTrials.gov Identifier: NCT01767311) was a multinational, multicenter, double-blind, placebo-controlled, parallel-group study (core) employing Bayesian design with response-adaptive randomization with an OLE (Figure
S1). Methods and primary results for the study 201 core phase have been published [
19]. Briefly, at entry into the core study, subjects were required to have early AD (amyloid positive) with global Clinical Dementia Rating (CDR) global score of 0.5 or 1. Subjects were randomized to either placebo or one of 5 active arms of lecanemab (2.5 mg/kg biweekly, 5 mg/kg monthly, 5 mg/kg biweekly, 10 mg/kg monthly, 10 mg/kg biweekly) without titration. Treatment duration of the study was 18 months with a 3-month follow-up and a target enrollment of approximately 800 subjects. The primary outcome was based on a Bayesian analysis at 12 months; the study continued per protocol with no unblinding to month 18. To maintain the blind during the double-blind portion of the trial, all subjects received biweekly infusions of either placebo or lecanemab.
The 201 OLE was initiated following analysis of the core phase 2b study 201 to allow subjects to receive open-label lecanemab 10 mg/kg biweekly (initiated without titration) for up to 24 months to assess long-term safety and tolerability. All subjects who fulfilled OLE inclusion/exclusion criteria and entered the OLE received 10 mg/kg biweekly during the OLE period. There was a gap period between the end of the study 201 core and OLE baseline when no treatment was provided. The gap period lasted for an average of 24 months (range 9–59 months) for all subjects who entered the OLE. Core treatment assignments remained blinded to study sites and study participants throughout the OLE.
Study assessments
Study assessments for the study 201 core and OLE included the Alzheimer’s Disease Composite Score (ADCOMS); Clinical Dementia Rating Sum-of-Boxes (CDR-SB); Alzheimer’s Disease Assessment Scale-Cognitive Subscale (ADAS-Cog14); changes in plasma biomarkers; and brain amyloid by PET Standardized Uptake Value ratio (SUVr) (in an optional substudy of consenting participants). The core amyloid PET substudy assessed baseline, 12-month, and 18-month SUVr with florbetapir; the OLE amyloid PET substudy assessments were at baseline, 3 or 6 months, 12 and 24 months. Plasma samples were collected at the same timepoints as the PET studies. Imaging (PET with florbetapir tracer) and plasma (amyloid-β (Aβ)42/40 ratio, phospho-tau (p-tau)181) biomarkers were evaluated. The amyloid PET SUVr normalized to whole cerebellum mask, measured using [
18]F florbetapir as a PET ligand, was used to determine brain amyloid levels [
20]. Plasma concentrations of Aβ42 and Aβ40 were measured using the immunoprecipitation/liquid chromatography-mass spectrometry/mass spectrometry (IP/LC-MS/MS) technology platform (Precivity AD assay, C2N), and the ratio of plasma Aβ42/40 was calculated from the output. Plasma concentrations of p-tau181 were measured using a commercially validated single molecule array (Simoa) assay developed by Quanterix. Information related to drug interference with plasma Aβ42/40 and p-tau181 assays can be found in the supplement.
Exposure response (ER) analyses for PET SUVr
The relationship between serum lecanemab concentration and the PET SUVr reduction time course in study 201 core and OLE phase was characterized by an indirect response model for the lecanemab concentration inducing the reduction of brain amyloid. Estimated parameters included baseline SUVr(t) at time=0 (BSUVr0), indirect response rate constant parameters (Kin and Kout [estimated as Kout = Kin/SUVr(0)]), maximum drug effect (Emax), and lecanemab concentration resulting in half of the maximum drug effect (EC50).
$$\frac{dSUVr}{dt}= Kin- SUVr(t)\ast Kout\ast \left[1+\frac{Emax\ast BAN2401\ Conc.}{EC50+ BAN2401\ Conc.}\right]$$
Inter-individual variability was estimated for baseline and Emax and could not be estimated for Kin and EC50. Residual variability was modeled using a proportional model. Covariates tested were sex (women vs men), age, neutralizing anti-drug antibodies (ADA) (positive vs negative), and apolipoprotein E4 (ApoE4) carrier status (positive vs negative only for Emax).
Exposure response analyses for efficacy
Longitudinal pharmacokinetic/pharmacodynamic (PK/PD) models were developed for efficacy endpoints (ADCOMS, CDR-SB, and ADAS-Cog) using data from the phase 2 study. A disease progression model was developed using data from placebo-treated subjects. Effect of model-predicted lecanemab exposure (maximum concentration at steady state [Css,max] and average concentration at steady state [Css,av]) on disease progression was investigated from data in all subjects as follows:
$$\textrm{EFF}-\textrm{INT}+\textrm{SLP}\ast \left(1-\textrm{DESLOPE}\ast \left(\textrm{lecanemab}\ \textrm{Exposure}\right)\right)\ast \textrm{Time}$$
where EFF, INT, SLP and DESLOPE are clinical scores of efficacy endpoints at each assessment time (EFF), baseline clinical score (INT), disease progression rate (SLP), and lecanemab effect on disease progression rate (DESLOPE), respectively.
Evaluated covariates were age, sex, ApoE4 carrier status (positive or negative), ongoing treatment with acetylcholinesterase inhibitors (AChEIs) and/or memantine (yes or no), and clinical subgroup (mild cognitive impairment (MCI) due to AD or mild AD dementia).
Relationship between PET SUVr and clinical efficacy
Relationships for change from baseline (CFB) of SUVr whole cerebellum (SUVrWC) versus CFB of clinical endpoints (ADCOMS, CDR-SB, and ADAS-Cog) at 12 and 18 months were explored in a subset of subjects who had post-baseline assessments for both endpoints. The relationships were modeled using a nonlinear effects model. A linear model (CFB of Clinical Endpoint = Intercept + Slope * CFB of PET SUVr) was explored for key clinical endpoints (ADCOMS, CDR-SB, and ADAS-Cog). For all endpoints, inter-individual variability (IIV) was estimated for intercept, and residual variability was modeled using an additive model. Effects of age, sex, ApoE4 carrier status, ongoing treatment with AChEIs and/or memantine (yes or no), and clinical subgroup (MCI due to AD or mild AD dementia) were evaluated as covariates. ER analyses for PET SUV efficacy were performed using nonlinear mixed-effect modeling in NONMEM® version 7.3. Where applicable, the final ER models were evaluated for performance using graphical assessment, non-parametric bootstrapping, and visual predictive checks.
Statistics
Statistical analyses for study 201 core have been previously published (Swanson 2021). In the study 201 OLE analyses, the focus is on de novo subjects (core placebo-treated) and those on 10mg/kg biweekly from beginning of study (delay start and early start design on most effective dosing regimen). Analyses were conducted in 2 cohorts based on their treatment allocation during study 201 core: (1) subjects who received prior placebo and (2) subjects with prior lecanemab 10 mg/kg biweekly. The change from OLE baseline in change in clinical endpoints (CDR-SB, ADCOMS, ADAS-Cog14) were analyzed using the mixed model repeated measures (MMRM) approach, incorporating key covariates into the model, ApoE4 status, clinical subgroup (MCI due to AD or mild AD dementia), ongoing treatment with AChEIs and/or memantine, and baseline value. Analyses of amyloid PET (SUVr and Centiloid approaches), plasma Aβ42/40, and plasma p-tau181 were also performed. Plasma biomarkers were measured for subjects with available samples. The correlations among the 3 biomarkers and their correlations with clinical endpoints were evaluated using population-level and subject-level correlation analysis. The OLE protocol was drafted and initiated after completion of the core study.
Discussion
This report describes the detailed results for the lecanemab study 201 core, gap, and OLE clinical and biomarker results. In addition, the relationships between clinical measures and amyloid PET imaging and soluble biomarkers were explored with correlations analyses. Lecanemab 10 mg/kg biweekly demonstrated the largest effect among tested doses on key biomarkers and clinical endpoints, reducing brain PET amyloid (measured by visual read, PET SUVr & Centiloid scale) with corresponding changes in plasma biomarkers, while slowing of clinical decline as measured by CDR-SB, ADCOMS, and ADAS-Cog14 was observed. There were consistent parallel directional relationships between biomarker changes and changes on clinical measures. Treatment with lecanemab 10 mg/kg biweekly results in a larger and faster decrease in amyloid PET SUVr, increase in plasma Aβ42/40 ratio (a more sensitive biomarker of amyloid cascade relative to PET SUVr in this study), and decrease in plasma p-tau181 as compared to lecanemab 10 mg/kg monthly dosing.
Lecanemab concentration was a significant predictor of brain amyloid removal in PK/PD (exposure –PET) modeling expressed as a maximum effect function (Emax) [
22]. Subjects with lower PET SUVr baseline achieved amyloid negativity faster than subjects with higher PET baseline values, but a higher baseline SUVr was associated with a greater magnitude of amyloid reduction. ApoE4 carrier status was identified as a significant covariate on baseline PET SUVr. ApoE4 carriers had higher baseline PET SUVr than ApoE4 noncarriers (1.39 vs 1.34). Age influenced maximum plaque removal independent of baseline PET SUVr level. For example, relative to a 72-year-old subject (median analysis set), an 84-year-old subject (95 percentile of analysis set) had 24% higher SUVr reduction, whereas a 57-year-old subject (5 percentile of analysis set) had 29% lower PET SUVr reduction. The half-life of brain amyloid re-accumulation as measured by amyloid PET was estimated to be approximately 4 years suggesting that it will take approximately 16 to 20 years (4–5 half-lives with an approximate half-life of 1.9 years for Aβ42/40 ratio degradation) for brain amyloid to reaccumulate and return to its value before lecanemab treatment.
Observed changes in plasma Aβ42/40 ratio during treatment discontinuation in the gap period suggest that stopping treatment leads to a reversal of the positive effects and at faster rate than Ab aggregation measured by PET. The plasma Aβ42/40 ratio begins decreasing again and plasma p-tau181 and amyloid PET SUVr to reverse their trajectory and start increasing, which are early indicators of brain amyloid accumulation [
23,
24], associated with clinical decline during the gap period. Parallel decline in the gap period between the treated group and the placebo group may suggest that continued treatment is needed to achieve a continuing therapeutic benefit. Initiation of lecanemab in the OLE reversed these negative biomarker trends. Continued treatment with lecanemab in the OLE showed continued improvement on multiple biomarkers used to track AD processes and considered signals reflecting the biology of AD. These findings suggest that continued targeting of protofibrils with lecanemab may be beneficial for patients while still in the early AD stage, even after brain amyloid clearance as measured by amyloid PET, because other forms of amyloid may exist that are not detected by amyloid PET. Therefore, continued dosing with lecanemab to a point of normalization of the plasma Aβ42/40 and p-tau181 levels may be necessary to better determine the disease-modifying effects of normalizing Aβ.
Clinical and cognitive outcomes during the gap and OLE periods were limited by low numbers and power in this analysis, and the benefits of the drug at later stages of disease are uncertain. Data from the ongoing phase 3 CLARITY OLE study will provide a better evaluation of the effect of high-dose lecanemab on cognitive and clinical outcomes at later stages of disease.
The results presented herein lead to several noteworthy conclusions. First, rapid and thorough amyloid reduction correlates with slowing of clinical decline. Lecanemab treatment can be initiated without titration with acceptable safety (Swanson 2021). Amyloid reduction is achieved within 3 months of treatment and clinical efficacy within 6 months of treatment, with >80% of subjects amyloid negative (by visual read) by 12–18 months. Finally, there may be potential to use plasma biomarkers to monitor lecanemab treatment effects. Correlations are observed among amyloid PET SUVr, clinical endpoints, and plasma biomarkers (Aβ42/40 ratio and p-tau181) following treatment with lecanemab. Monitoring of treatment effects using plasma biomarkers may allow dose modification as needed following rapid and pronounced amyloid removal (e.g., less frequent and/or lower dose). This may obviate the need for repeat of PET scans to determine amyloid status, a current limitation in the delivery of this class of therapies to a broader population. However, the data on the effects of longer-term dosing with lecanemab on plasma biomarkers is needed to better determine the true potential of these biomarkers in monitoring therapeutic response over the long term. The more rapid return towards disease levels of plasma biomarkers relative to amyloid PET during the gap period suggests that the plasma measures may be a more dynamic measure of disease state to determine chronic dosing strategies after amyloid PET levels are normalized.
Anti-amyloid monoclonal antibodies were developed based on the central role of amyloid in AD and the hypothesis that decreasing fibrillar and protofibrillar amyloid would lead to disease modification and slowing of cognitive decline. There is not a consensus on the data needed to conclude that an agent is disease-modifying. However, three features consistent with disease modification were observed in this trial: (1) no return to the placebo level of the treated participants with cessation of therapy; (2) effects on biomarkers (Aβ, p-tau) considered important features of fundamental AD biology; and (3) persistent change in the trajectory of the illness that, generally, correlate with disease biomarkers and that supports modification of the underlying pathophysiology of the disease [
25‐
27]. These features will contribute to the data accumulating to support the potential disease-modifying effects of lecanemab. Moreover, the temporal relationship between the soluble biomarkers and aggregated amyloid PET during the core, gap, and OLE phase provides unique information on the effects of amyloid reduction with anti-Aβ monoclonal antibodies, specifically recapitulating the sequence of events reported in observational studies. Although there have been a number of recent reports to suggest amyloid plaques are associated with the initial rise in soluble p-tau and that the reduction of amyloid plaques with anti-Aβ monoclonal antibodies results in a reduction of soluble p-tau biomarkers, the gap period in this study suggests that when anti-Aβ monoclonal antibodies are discontinued soluble amyloid (in the form of the plasma Aβ42/40 ratio) begins to return towards baseline levels, followed by plasma p-tau and both clearly precede the slow re-accumulation of amyloid PET, similar to what has been observed in observational studies [
28‐
30]. This indicates that at the cessation of treatment, even with very low amyloid PET values, there is likely a reaccumulating of Aβ aggregates leading to the rise of both soluble Aβ and p-tau.
In addition to lecanemab, several anti-Aβ monoclonal antibodies with distinct Aβ binding profiles (e.g., aducanumab, bapineuzumab, gantenerumab, and solanezumab) have emerged and are in various stages of clinical development [
27,
31,
32]. All these potential therapies are based on disease models which suggest that tau pathology is triggered by Aβ, leading to AD progression [
33]. Although most previously published studies using other putative disease-modifying agents for AD did not show appreciable clinical efficacy in phase 3 [
27], several recent studies have shown promising effects on reducing brain amyloid levels and slowing clinical decline [
19,
34,
35]. The lecanemab mechanism of action is distinct among other anti-amyloid agents. Lecanemab has high selectivity for soluble aggregated species of Aβ compared to monomeric amyloid, with moderate selectivity over fibrillar amyloid, a profile thought to convey an advantage in selectively targeting the most toxic pathologic amyloid species [
12,
15,
19].
There are several limitations of this analysis. Of the 856 randomized subjects, 180 voluntarily enrolled into the OLE. Thus, subjects were not randomized by treatment and key disease characteristics into the OLE. In addition, the OLE was started after a delay, resulting in a variable length gap period ranging from 9 to 59 months. If the OLE had started immediately after the core phase of study 201, more information on continuous dosing could have been obtained. However, the gap period presented the opportunity to observe subjects when anti-amyloid therapy was interrupted (i.e., the gap between the core and OLE) and then restarted in the OLE.
Declarations
Competing interests
EM is the Associate Director of the DIAN–TU. He reports serving on a Data Safety Committee for Eli Lilly and Company and Alector. He is scientific consultant for Eisai and Eli Lilly and Company and has received institutional grant support from Eli Lilly and Company, F. Hoffmann-La Roche Ltd. and Janssen. JC has provided consultation to Acadia, Actinogen, Alkahest, AlphaCognition, AriBio, Biogen, Cassava, Cerecin, Cortexyme, Diadem, EIP Pharma, Eisai, eqt, GemVax, Genentech, Green Valley, GAP Innovations, Grifols, Janssen, Karuna, Lilly, Lundbeck, Merck, NervGen, Novo Nordisk, Oligomerix, Optoceutics, Ono, Otsuka, PRODEO, Prothena, ReMYND, Resverlogix, Roche, Sage Therapeutics, Signant Health, Suven, TrueBinding, and Vaxxinity pharmaceutical, assessment, and investment companies. JC is supported by NIGMS grant P20GM109025; NINDS grant U01NS093334; NIA grant R01AG053798; NIA grant P20AG068053; NIA grant P30AG072959; NIA grant R35AG71476; Alzheimer’s Disease Drug Discovery Foundation (ADDF); Ted and Maria Quirk Endowment; and the Joy Chambers-Grundy Endowment. SD, CS, LR, MK, AK, MI, and LK are employees of Eisai. RJB is Director of DIAN–TU and Principal Investigator of DIAN–TU-001. He receives research support from the NIA of the NIH, DIAN–TU trial pharmaceutical partners (Eli Lilly and Company, F. Hoffman-La Roche Ltd and Avid Radiopharmaceuticals), Alzheimer’s Association, GHR Foundation, Anonymous Organization, DIAN–TU Pharma Consortium (active: Biogen, Eisai, Eli Lilly and Company, Janssen, F. Hoffmann-La Roche Ltd/Genentech; previous: AbbVie, Amgen, AstraZeneca, Forum, Mithridion, Novartis, Pfizer, Sanofi, United Neuroscience). He has been an invited speaker and consultant for AC Immune, F. Hoffman-La Roche Ltd and Janssen and a consultant for Amgen and Eisai. The author(s) read and approved the final manuscript
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.