Abstract Body

Plasma assays for amyloid beta, phosphorylated tau, neurofilament light and other markers are now the most innovative and exciting development in the field of diagnostics for Alzheimer’s Disease (AD). The Alzheimer’s Disease Neuroimaging Initiative is in an ideal position to help validate and compare these assays, because plasma samples have been banked on over 2,270 participants at each longitudinal visit. ADNI data on plasma assays will be discussed.

The goal of ADNI (see ADNI-info.org) has been to standardize and validate MRI and PET imaging and blood/CSF biomarkers for Alzheimer’s treatment trials. Overall, we have enrolled a total of 2,272 participants: Elders with mild cognitive impairment (MCI) (n=1043); Dementia (n=397); Normal Controls (n=518) and 301 Controls with cognitive complaints. Participants have annual clinical visits, neuropsychological assessments, MRI (structural, perfusion, diffusion-tensor and task-free, resting-state fMRI), FDG PET, both amyloid and tau PET, blood and urine, and CSF (abeta, tau, ptau and other analytes), and whole genome sequencing. All ADNI data is available to all scientists in the world, on USC/LONI/ADNI, without embargo. A major interest in the field is to identify predictors of future risk for cognitive decline, and ADNI provides the largest available data set for analyses of this type. Over 2,200 papers have been published on ADNI.

Plasma, DNA, RNA, RBC pellets, and CSF samples can be requested online at http://adni.loni.usc.edu/data-samples/access-data/

ADNI will validate diagnostic techniques, including the new plasma assays, which can be used for diagnosis, and to measure the effects of treatments which slow progression and prevent AD.

Aims

Assess the associations of plasma neurofilament light polypeptide (NFL) and tau phosphorylated at threonine 181 (pTau) with cerebral AD pathology and with cognitive decline in a memory clinic setting.

Methods

Two hundred eighteen elderly subjects with normal cognition (NC, n=90) and cognitively impaired memory clinic patients (CI: mild cognitive impairment (n=57) or dementia (n=71)) were included. Concentrations of NFL and pTau and CSF biomarkers of cerebral AD pathology were measured. Decline in global cognition and progression of disease severity were assessed by changes in MMSE and CDRSoB, at follow-up visits (FU). Multivariate analysis assessed associations of NFL and pTau with AD pathology, single CSF and brain volumetry biomarkers, and clinical disease progression.

Results

Plasma NFL and pTau levels were associated with AD pathology in both groups, and correlated with CSF biomarkers and regional brain volumetry . After controlling for covariates, pTau levels remained associated with AD in CI participants. Adding pTau to a reference model improved prediction of cerebral AD pathology in CI participants (AUC 0.861). In NC plasma NFL predicted changes in CDRSoB while in CI participants. NFL predicted CDRSoB change and pTau predicted changes in bothe MMSE and in CDRSoB scores. After controlling for covariates, only the association NFL with CDRSoB in NC participants remained significant. Adding pTau to a reference model improved MMSE change prediction in CI participants (AUC 0.838).

Conclusions

NFL and pTau levels can serve as biomarkers of cerebral AD pathology and cognitive decline. Their predictive performance depends on the presence of cognitive impairment.

Abstract Body

The Aβ-misfolding is a promising structure biomarker which could identify AD up to 14 years before clinical conversion ^1,2,3,4. Additional use of the Tau structure biomarker increases the sensitivity and specificity⁵.

Objectives:

The performance as a prognostic plasma biomarker in individuals with subjective cognitive decline (SCD) is shown⁷.

Methods:

Baseline plasma samples of SCD subjects were analyzed using the immuno-infrared-sensor. We used COX proportional hazard models to quantify the Ab-misfolding as prognostic biomarker for future clinical conversion to mild cognitive impairment (MCI) or Alzheimer’s disease (AD). The accuracy was determined by time-dependent ROC-curve-analyses.

Results:

A hazard ratio (HR) of 19 was obtained; T-ROC curve analyses yielded an AUC of 0.94. Using in addition plasma Ab42/40 provides an added value with an AUC of 0.99 six years before clinical conversion^6,7.

Conclusions:

The proposed plasma biomarker panel can precisely predict conversion to clinical MCI and AD in cognitively unimpaired subjects. They allow screening of the aging population for prevention and early intervention in symptom-free stages.

1. Nabers A, et al. J.Biophotonics. 2016;9(3):224-34.

2. Schartner J, et al. ACSMed.Chem.Lett. 2017Jun11;8(7):710-714.

3. Nabers A, et al. EMBO.Mol.Med. 2018May;10(5).

4. Stocker H, et al. Alzheimer&Dement. . 2020;16:283–91.

5. Nabers A, et al. Alzheimer&Dement. 2019;11:257-263.

6. Verberk et al. Ann Neurol. 2018;84:648–58.

7. Stockmann J*,Verberk, I…, Scheltens, P. Teunissen CE, Gerwert K. under review.

Abstract Body

Recent developments in characterizing tau species by Mass Spectrometry in the human brain, cerebrospinal fluid (CSF) and plasma have provided new insights into the diversity of tau isoforms and their changes associated with Alzheimer disease (AD) over the course of the disease. Quantitation of CSF phosphorylated tau (p-tau) in familial and sporadic AD has highlighted T217 site as an accurate biomarker in identifying asymptomatic amyloid positive individuals, well before the development of tau aggregates. P-tau217 outperforms the well-established p-tau181 as biomarker for AD diagnosis in both CSF and plasma. This suggests that plasma p-tau217 may serve as the most accurate target for AD diagnosis using a blood-based assay. Additional multiple CSF p-tau species exhibit different behaviors over the course of the disease, together providing more accurate and specific biomarkers to stage AD. These p-tau species may differently be impacted by pathologic metabolic changes and tau aggregation in the brain. Indeed, quantitative measures of multiple brain p-tau species in AD tau aggregates suggest site-specific enrichment of specific p-tau isoforms.

Furthermore, CSF truncated tau species containing the microtubule binding region (MTBR), the domain primarily accumulated in AD tau aggregates, demonstrate promising clinical utility to monitor tau pathology. These peptides are enriched in brain insoluble tau and corresponding CSF levels correlate with brain tau pathology measured by PET.

In conclusion, the accurate measures of tau post-translational modifications of tau can serve as new biomarkers to complement established p-tau and tau assays and refine AD diagnosis and staging.

Aims

It is currently unclear whether plasma biomarkers can be used as prognostic tools for Alzheimer’s disease (AD). Here we address this question by examining plasma amyloid-β 42/40 (Aβ42/40), phosphorylated-tau 181 (P-tau181), phosphorylated-tau 217 (P-tau217) and neurofilament light (NfL) in non-demented individuals who underwent longitudinal amyloid and tau positron emission tomography (PET), magnetic resonance imaging (MRI) and cognitive testing.

Methods

Blood was collected at baseline from all participants to determine the levels of Aβ42/40, P-tau181, P-tau217 and NfL. All subjects underwent longitudinal ¹⁸F-RO948 PET, structural MRI and cognitive assessment. In addition, a subsample also had longitudinal ¹⁸F-flutemetamol PET scans. Linear mixed effects models and voxel-wise analyses were applied to assess the relationship between plasma biomarkers and longitudinal changes in imaging and cognition.

Results

Our results show that plasma P-tau217 predicts increases in amyloid (p=0.017) and tau (p=0.004) PET, medial temporal atrophy (p<0.001) and global cognitive decline (p=0.005). The results for the other plasma markers were more variable, with plasma Aβ42/40 predicting amyloid (p=0.002) and tau (p=0.003) PET signal increases as well as cognition (p=0.020), whereas P-tau181 predicts hippocampal atrophy (p=0.008) and cognitive decline (p=0.003), and finally NfL predicts atrophy (p=0.001) and tau PET increases (p=0.020). The comparison between the models showed that the ones that included P-tau217 as a predictor had a significantly better fit to the data as indicated by lower Bayesian Information Criterion values.

Conclusions

These findings suggest that plasma P-tau217 could be useful in clinical trials to determine whether an individual is on the pathophysiological pathway of AD.

Aims

Plasma p-tau181 and p-tau217 have shown excellent utility in the differential diagnosis of dementia patients. However, increases at the preclinical stage are subtle and may not be optimal for detecting preclinical pathology. Our recent work in cerebrospinal fluid (CSF) indicates p-tau231 to be a preclinical biomarker for AD pathology, which motivated us to develop a novel molecule array (Simoa) to detect p-tau231 in plasma.

Methods

We developed and validated a novel plasma p-tau231 in-house Simoa assays at the Clinical Neurochemistry Laboratory, Sahlgrenska University Hospital, Sweden, in 541 participants of 3 independent cohorts.

Results

In the discovery cohort (n=38), plasma p-tau231 demonstrated high accuracy to predict AD (AUC=0.941). In the validation cohort (n=313), plasma p-tau231 had high accuracy in determining AD from young (AUC=0.951), Aβ- elderly (AUC=0.915), Aβ- MCI (AUC=0.883) and other neurodegenerative disorders (AUC=0.923). In addition, plasma p-tau231 showed remarkable accuracy for the identification of tau PET and Aβ PET positivity (AUC=0.93). Comparing Aβ- elderly and Aβ+ elderly individuals without cognitive decline, p-tau231 had a high accuracy (AUC=0.884) to detect preclinical Aβ pathology, which was superior to plasma p-tau181 (AUC=0.771). Plasma p-tau231 predicted longitudinal cognitive decline (P=0.02) and correlated with Aβ PET, tau PET and CSF p-tau231 (r>0.610).

Conclusions

Our results indicate that plasma p-tau231 might be a superior marker for identifying preclinical AD. Thus, plasma p-tau231 has the potential to provide an invaluable contribution as a biomarker of both screening of preclinical AD and AD dementia diagnosis for the use in clinical practice and trials.

Aims

1) Utilize the NIA-AA research framework to characterize the frequency of cerebrospinal fluid (CSF) amyloid (A) and tau (T) positivity (-/+) across the Alzheimer’s disease (AD) clinical spectrum; 2) among unimpaired A/T groups (A-T-, A-T+, A+T-, A+T+), examine cognitive trajectories and associations with CSF analytes, including candidate biomarkers for neurodegeneration.

Methods

Individuals (N=642) recruited from Wisconsin AD Research Center (WADRC, n=404) and Wisconsin Registry for Alzheimer’s Prevention (WRAP, n=238) studies were assayed using NeuroToolKit (NTK) robust prototype assays (Roche Diagnostics). Differences among A/T groups (demographics, cognition, and NTK markers of degeneration) were assessed using ANOVA or χ² tests. In subsets unimpaired at baseline (WADRC, n=302; WRAP, n=232), longitudinal PACC-3 performance (Num. assessments, Median[Range]=5[2,9]) across A/T biomarker groups was investigated (linear mixed effects models).

Results

Amyloid positivity (A+) increased with clinical status (unimpaired: 19%, MCI: 63%, dementia: 92%). Among unimpaired (n=533), A+ frequency was higher in subjects aged >65 years and among APOE4+. On NTK measures, unimpaired A/T groups differed on neurodegeneration (4/4 analytes) and glial activation (3/4), but not inflammation (0/1). Neurodegeneration analyte levels were higher in T+ groups; neurogranin demonstrated a stepwise significance pattern of A-T- < A+T- < both A-T+ and A+T+. Although A/T group baseline cognition did not differ, the A+T+ and A+T- groups exhibited faster cognitive decline versus A-T-.

Conclusions

In two separate preclinical cohorts, cognitive decline was evident in A+ biomarker groups. Several NTK analytes, particularly neurogranin, were sensitive to A and/or T positivity indicating that these analytes differ preclinically and require further study.

Aims

To evaluate how plasma glial fibrillary acidic protein (GFAP) associates with core PET, CSF and plasma biomarkers of the Alzheimer’s disease (AD) pathophysiology.

Methods

Plasma GFAP was quantified using a commercial Simoa assay on 330 participants of the Translational biomarker in aging and dementia (TRIAD) cohort (36 young individuals (<30 years old), 168 cognitively unimpaired (CU), 62 mild cognitive impairment (MCI), 45 AD, and 12 frontotemporal dementia (FTD) patients). These participants had also [18F]AZD4694 and [18F]MK6240 PET imaging data. Participants had also quantified CSF Ab42/40, p-tau181 and t-tau on the Lumipulse platform. In addition, CSF and plasma neurofilament (NfL) light and plasma p-tau181 were quantified with in house assays on the Simoa platform. The distribution of plasma GFAP concentration was compared across groups using linear regression models adjusting for age and sex. In addition, Spearman rank correlations were employed between biomarkers. Finally, mediation analyses were used to further evaluate the relationship between plasma GFAP and amyloid and tau pathologies in this context.

Results

Plasma GFAP was found to progressively increase with age (P=3.4x10^-9) and disease severity (P=4.6x10^-11), showing its lowest concentration levels in young participants and the highest in AD patients (Figure 1). Plasma GFAP was also highly correlated with amyloid pathology, neurodegeneration biomarkers and tau pathology. Interestingly, mediation analysis indicated that the associations with t-tau, p-tau181 and tau PET biomarkers are dependent on amyloid pathology levels.

Conclusions

Plasma GFAP is a biomarker that reflects amyloid pathology in the AD spectrum.