Polygenic hazard score: an enrichment marker for Alzheimer's associated amyloid and tau deposition

Abstract

There is an urgent need for identifying nondemented individuals at the highest risk of progressing to Alzheimer's disease (AD) dementia. Here, we evaluated whether a recently validated polygenic hazard score (PHS) can be integrated with known in vivo cerebrospinal fluid (CSF) or positron emission tomography (PET) biomarkers of amyloid, and CSF tau pathology to prospectively predict cognitive and clinical decline in 347 cognitive normal (CN; baseline age range = 59.7-90.1, 98.85% white) and 599 mild cognitively impaired (MCI; baseline age range = 54.4-91.4, 98.83% white) individuals from the Alzheimer's Disease Neuroimaging Initiative 1, GO, and 2. We further investigated the association of PHS with post-mortem amyloid load and neurofibrillary tangles in the Religious Orders Study and Memory and Aging Project (ROSMAP) cohort (N = 485, age at death range = 71.3-108.3). In CN and MCI individuals, we found that amyloid and total tau positivity systematically varies as a function of PHS. For individuals in greater than the 50th percentile PHS, the positive predictive value for amyloid approached 100%; for individuals in less than the 25th percentile PHS, the negative predictive value for total tau approached 85%. High PHS individuals with amyloid and tau pathology showed the steepest longitudinal cognitive and clinical decline, even among APOE ε4 noncarriers. Among the CN subgroup, we similarly found that PHS was strongly associated with amyloid positivity and the combination of PHS and biomarker status significantly predicted longitudinal clinical progression. In the ROSMAP cohort, higher PHS was associated with higher post-mortem amyloid load and neurofibrillary tangles, even in APOE ε4 noncarriers. Together, our results show that even after accounting for APOE ε4 effects, PHS may be useful in MCI and preclinical AD therapeutic trials to enrich for biomarker-positive individuals at highest risk for short-term clinical progression.

Keywords: Alzheimer’s disease; Amyloid; Polygenic risk; Tau.

Fig. 1

Flowchart delineating the hypotheses tested in the a Alzheimer’s disease neuroimaging initiative (ADNI) cohort and b religious orders study and memory and aging project (ROSMAP) cohort

Fig. 2

Increase in the proportion of nondemented individuals who tested positive on amyloid (red) and total tau (blue) as a function of higher polygenic hazard score (PHS) (moving 10 percentiles, 1% increment per step)

Fig. 3

a Positive predictive value (PPV) and b negative predictive value (NPV) of amyloid and total tau with subsequent progression to AD dementia, based on stratification of polygenic hazard score (PHS) into different percentile bins (≤ 25%, all individuals, ≥ 50% and ≥ 75%). Error bars are 1000 bootstrap estimate of the standard deviation of 100 random samples in each PHS percentile bins

Fig. 4

Differences in rates of cognitive decline in a executive function, b memory, and clinical progression in c cognitive dementia rating sum of boxes (CDR-SB) over time for high polygenic hazard score (PHS) individuals who tested positive for amyloid (solid red line) or total tau (solid blue line) in full PHS linear mixed-effects (LME) models, compared to individuals who tested positive for amyloid (dotted red line) or total tau (dotted blue line) in reduced non-PHS LME models (see text for model details)

Fig. 5

Survivor plot showing greater progression to AD dementia as a function of polygenic hazard score (PHS) for all individuals (green), individuals who were amyloid positive (red) and individuals who were both amyloid and total tau positive (black)

Fig. 6

Annualized cumulative incidence rates depicting instantaneous hazard based on an individual’s age and polygenic hazard score (PHS), stratified based on individuals who were amyloid positive (red), amyloid negative (orange), total tau positive (blue), total tau negative (purple), and both amyloid and total tau positive (black)