Aims

Identification of early amyloid pathology can be of value for preventive trials. We assessed whether global and regional Centiloid (CL) predicted conversion to amyloid-positivity based on PET visual read (VR) in a cognitively unimpaired (CU) population.

Methods

Longitudinal [¹⁸F]flutemetamol amyloid-PET scans were performed on a Philips PET-MRI system in 140 CU participants from the EMIF-AD Twin60++ study (Table 1), with an average follow-up (FU) time of 4.34y (3.7–6.2). All scans were visually assessed as negative or positive (VR-/VR+) and subjects were classified as Stable VR-, Converter or Stable VR+. Global and regional quantitative values were obtained using the CL method. Generalized Estimating Equation (GEE) was used to assess differences between VR-groups in global and regional CL baseline burden and yearly rate of change (ΔCL). Analyses were corrected for age, sex, and, APOE-ε4 carriership.

Results

At baseline, 14.3% of subjects were considered VR+ and displayed higher CL burden compared to VR- cases (β=39.6, p<.001). At FU, 17 (14.2%) subjects converted to VR+. Converters did not have higher baseline global CL burden, but did show a higher global ΔCL compared to stable VR- cases (β=3.3, p<.001), Figure 1A). In contrast, the PC/PCC ROI of converters had both a higher baseline CL (β=20.9, p<0.001 ) and ΔCL compared to Stable VR- subjects (β=5.0, p<.001, Figure 1B, Table 2).

Conclusions

Baseline regional CL quantification in the PC/PCC, rather than global, could predict conversion to amyloid-positivity based on visual assessment in a CU population. This could support selection of participants for future primary prevention trials.

Aims

The amyloid-tau-neurodegeneration (AT(N)) Alzheimer’s disease (AD) research framework has been increasingly expanded to include other biomarkers, such as cerebrovascular disease (V). We characterized cortical thickness by ATV profiles in a racially and ethnically diverse, middle-aged community-based cohort.

Methods

Participants (n=65, age=61±6, sex=68% women, race=11% Non-Latinx White, 25% Non-Latinx Black, 64% Latinx, APOE4=32% carriers) were from the Offspring Study of Racial and Ethnic Disparities in Alzheimer’s Disease. Amyloid and tau positivity were determined in Thal and Braak regions, respectively; 2SD above the mean was used as the positivity threshold. Total white matter hyperintensity volume, representing ischemic small vessel disease, was the primary marker of cerebrovascular disease; the median volume was used as the positivity threshold. Cortical thickness in the AD signature was compared across ATV profiles.

Results

Half (48%) of middle-aged adults had no pathology (A-T-V-) based on our positivity thresholds, 37% had vascular pathology only (A-T-V+), 5% had AD pathology only (A+T-V-, A+T+V-), 9% had both AD and vascular pathology (A+T-V+, A+T+V+), and one person had non-AD and vascular pathology (A-T+V+; excluded from subsequent analyses). Compared to those with no pathology, AD signature was numerically lower in those with vascular pathology only (-0.01 [-0.06, 0.04]) and in those with AD pathology with and without vascular pathology (-0.03 [-0.10, 0.03]).

Conclusions

Cerebrovascular pathology was the most common midlife pathology. AD without cerebrovascular pathology was less common than AD with cerebrovascular pathology. Progressive ATV profiles were descriptively related to lower cortical thickness, but should be confirmed as more participants are imaged.

Aims

Alzheimer’s disease (AD) pathophysiology is heterogenous. We studied if distinct subtypes could be detected with cerebrospinal fluid (CSF) proteomics.

Methods

We selected 419 AD individuals (with an abnormal amyloid biomarker) and 187 controls (with normal cognition and normal AD biomarkers) from Alzheimer center Amsterdam studies. With 16-plex TMT MS technology we detected 3987 proteins in CSF, of which 1310 proteins were observed in everyone and used for further analyses. We clustered 1059 proteins associated with AD (p<.05). Potential upstream drivers of molecular subtypes were identified with ENRICHR.

Results

We found 5 subtypes with distinct protein profiles: Compared to controls, 3 subtypes(1a, 1b, 1c) had high levels of neuronal plasticity proteins, and two(2a, 2b) had low levels of plasticity proteins. 137(32%) subjects had subtype 1a, with increased amyloid processing with high BACE1 and abeta40 levels. 560 subtype 1a proteins associated with REST and SUZ12 as potential upstream drivers. 124(30%) subjects had subtype 1b, with high levels of inflammation proteins and complement activation. 593 subtype 1b proteins converged on SOX2 as potential upstream driver. 24(6%) subjects had subtype 1c with high levels of 154 proteins converging on TCF3 as potential upstream driver. Subtypes 2a included 78(19%) individuals with high levels of 318 proteins expressed by the choroid plexus that converged on NFE2L2. Subtype 2b included 56(13%) subjects with high levels of 327 proteins associated with blood-brain barrier leakage, converging on ERG1.

Conclusions

We found three AD subtypes showing amyloid overproduction, and two with impaired amyloid clearance. This underlines the necessity of personalized treatment in AD.

Aims

Alzheimer’s disease (AD) is the most common cause of dementia, affecting over 50 million people globally. Patient stratification may be used to identify individuals who may experience greater benefit from treatment options, particularly with emerging disease-modifying treatments (DMTs). Here, we discuss the importance, challenges, and evolution of patient stratification in AD.

Methods

In January 2021, a panel of clinicians and AD experts met to discuss the current application of patient stratification in AD. Subsequently, a qualitative literature review was conducted for highly important topics in stratification.

Results

Vital applications for patient stratification include differential diagnosis of AD from other dementias, staging of disease severity, identifying people at risk of clinical progression or amyloid-related imaging abnormalities, tailoring individual treatment pathways, and determining eligibility for clinical trial participation. A variety of methods are used for stratification according to demographics, vascular risk factors, genotyping, and symptomatology (cognitive and functional assessments). AD pathology biomarkers available (amyloid or tau assessed by positron emission tomography or cerebrospinal fluid ascertainment) can also help in patient stratification; but availability and application vary across countries, hampering correct identification of patients at risk. Similarly, stigma for patients or clinicians, misdiagnosis, and poor coordination of the diagnostic process between primary care and memory clinics challenge utilisation of patient stratification. Moreover, current diagnostic testing may be cost-prohibitive or perceived as invasive.

Conclusions

Despite major challenges, patient stratification is an important tool in AD treatment that can help determine which individuals would benefit from treatment, especially as DMTs become more widely available.

Abstract Body

We verified the performance of Abeta misfolding as a prescreening plasma biomarker for AD as recently shown [1,2,3] with an extended observation period up to 17 years before clinical onset. Additionally, the structure biomarker was compared to the concentration biomarkers GFAP, NFL and ptau181 [4].

Baseline plasma samples of 308 subjects taken between 2000-2002 were analyzed using the infrared-immuno-sensor (iRIS) [3]. The obtained structure biomarker results were compared with GFAP, NFL and ptau181 levels measured with SIMOA.

All biomarkers showed significant alterations in the 17 year old baseline samples of later AD patients compared to the healthy control subjects [4]. However, the structure biomarker showed the best prognostic performance at 17-year follow-up relative to all concentration biomarkers. Additionally, a biomarker panel of the structure biomarker and GFAP levels showed an added value. Interestingly, the prognostic performance of ptau181 was limited to about 8 years before symptom onset. It could not predict AD conversion much more than 8 years in advance.

The structure plasma biomarker identifies individuals with high risk to develop AD up to 17 years before clinical onset [4]. This allows screening by a simple blood test of the aging population for prevention and early intervention in symptom-free stages.

1. Nabers A, et al. EMBO Mol. Med. 2018

2. Stockmann J and Verberk I, et al. Alz Res and Ther. 2020

3. Nabers A, et al. J. Biophotonics. 2016

4 . Léon Beyer and Hannah Stocker, Dan Rujescu, Bernd Holleczek, Julia Stockmann, Andreas Nabers, Hermann Brenner, Klaus Gerwert. submitted

Abstract Body

Current theories of Alzheimer’s disease (AD) pathogenesis postulate that β-amyloid facilitates spread of pathological tau from the entorhinal cortex (ERC) to neocortically connected targets. Much remains unclear about this model, including the role of neural activity. We studied normal older participants from the Berkeley Aging Cohort Study using PET measures of β-amyloid and tau, and functional MRI. We investigated relationships between functional connectivity, neural activity during performance of a memory task, and patterns and rates of tau deposition. First, we show that functional connectivity between the hippocampus and medial parietal cortex (MPC) is related to tau deposition in MPC, thus providing evidence that the pathway from ERC to neocortex transits via the hippocampus. We also show that neural activity in the medial temporal lobe (MTL) at baseline is related to the rate of subsequent tau deposition. Higher activation in ERC and parahippocampal cortex during a memory task in ERC is related to faster tau accumulation in these regions. In order to investigate these effects on MTL function, we examined the phenomenon of repetition suppression, whereby neural activity is normally suppressed for the repeated presentation of a previous stimulus. We found that tau was associated with increased neural activity during the memory task, and was also associated with reduced repetition suppression. Together these findings support a model whereby tau spreads to neocortex via the hippocampus, with higher neural activity related to both higher rates of tau deposition and dysfunction of the MTL memory system.

Abstract Body

Alzheimer’s disease (AD) is a remarkably heterogenous disorder in terms of clinical and cognitive presentations, longitudinal trajectories and anatomic patterns of neurodegeneration. While most patients present with progressive memory loss, a significant minority show primary decline in language (i.e., logopenic variant of primary progressive aphasia), visuospatial/visuoperceptual abilities (posterior cortical atrophy), executive functions, behavior or sensorimotor integration. This talk will describe the neuroimaging correlates of heterogeneity in AD, evaluating in vivo the relationships between the burden and distribution of amyloid and tau, neurodegeneration, brain physiology and clinical features of the disease. We will present how data-driven approaches have revealed reproducible patterns of tau deposition and neurodegeneration in AD, each corresponding to unique patient characteristics and clinical profiles. We will review putative modulators of AD phenotypes, including age at symptom onset, sex, APOE (and other genetic factors), neurodevelopmental differences and comorbid neuropathologies. We will conclude by discussing ramifications of heterogeneity for drug development and future precision medicine approaches for treating AD.