The dynamics of plasma biomarkers across the Alzheimer's continuum

Abstract

Background: Failures in drug trials strengthen the necessity to further determine the neuropathological events during the development of Alzheimer's disease (AD). We sought to investigate the dynamic changes and performance of plasma biomarkers across the entire Alzheimer's continuum in the Chinese population.

Methods: Plasma amyloid-β (Αβ)42, Aβ40, Aβ42/Aβ40, phosphorylated tau (p-tau)181, neurofilament light (NfL), and glial fibrillary acidic protein (GFAP) were measured utilizing the ultrasensitive single-molecule array technology across the AD continuum (n=206), wherein Aβ status was defined by the values of cerebrospinal fluid (CSF) Aβ42 or Aβ positron emission tomography (PET). Their trajectories were compared with those of putative CSF biomarkers.

Results: Plasma GFAP and p-tau181 increased only in Aβ-positive individuals throughout aging, whereas NfL increased with aging regardless of Aβ status. Among the plasma biomarkers studied, GFAP was the one that changed first. It had a prominent elevation early in the cognitively unimpaired (CU) A+T- phase (CU A+T- phase: 97.10±41.29 pg/ml; CU A-T- phase: 49.18±14.39 pg/ml; p<0.001). From preclinical to symptomatic stages of AD, plasma GFAP started to rise sharply as soon as CSF Aβ became abnormal and continued to increase until reaching its highest level during the AD dementia phase. The greatest slope of change was seen in plasma GFAP. This is followed by CSF p-tau181 and total-tau, and, to a lesser extent, then plasma p-tau181. In contrast, the changes in plasma NfL, Aβ42/Aβ40, Aβ42, and Aβ40 were less pronounced. Of note, these plasma biomarkers exhibited smaller dynamic ranges than their CSF counterparts, except for GFAP which was the opposite. Plasma GFAP and p-tau181 were tightly associated with AD pathologies and amyloid tracer uptake in widespread brain areas. Plasma GFAP could accurately identify CSF Aβ42 (area under the curve (AUC)=0.911) and Aβ PET (AUC=0.971) positivity. Plasma p-tau181 also performed well in discriminating Aβ PET status (AUC=0.916), whereas the discriminative accuracy was relatively low for other plasma biomarkers.

Conclusions: This study is the first to delineate the trajectories of plasma biomarkers throughout the Alzheimer's continuum in the Chinese population, providing important implications for future trials targeting plasma GFAP to facilitate AD prevention and treatment.

Keywords: Alzheimer’s continuum; Biomarker; Glial fibrillary acidic protein; Plasma; Trajectory.

Fig. 1

Group comparisons of plasma biomarkers. Plasma levels of GFAP, p-tau181, Αβ42/Αβ40, NfL, Αβ42, and Αβ40 between groups were compared with one-way analysis of covariance controlling for age and sex, followed by FDR corrected pair-wise post hoc comparisons. Significance: ****p<0.0001, ***p<0.001, **p<0.01, *p<0.05, −: p≥0.05. Abbreviations: GFAP, glial fibrillary acidic protein; p-tau, phosphorylated tau; Αβ, amyloid-β; NfL, neurofilament light; FDR, false discovery rate; MCI, mild cognitive impairment; AD, Alzheimer’s disease

Fig. 2

Correlations of plasma markers with demographic features. Demographic factors (age, sex, education, and APOE ε4 status) correlated to plasma measures (log-converted) were analyzed using multiple linear regressions. When one of the demographic factors was analyzed as a predictor variable, the other three factors were included in the model as covariates. The dotted gray lines from the center to the periphery represented the derived correlation coefficient (β) ranging from −0.080 to 0.405 in the CU A−T− and the whole Alzheimer’s continuum (CU A+T−, CU A+T+, MCI+, and AD dementia) groups (A) and from -0.161 to 0.271 in the CU A−T− and preclinical AD (CU A+T−, CU A+T+) groups (B). The * indicated the FDR-corrected p<0.05. C Scatter plots showing the associations of each plasma biomarker with age in the CU A−T− and preclinical AD groups. The standardized regression coefficients (β) and the P-values were computed using multiple linear regressions adjusting for sex, education, and APOE ε4 status. Additionally, we calculated the “age×Aβ status” interaction term. All P-values are corrected for multiple comparisons using the FDR approach. Abbreviations: APOE, apolipoprotein E; MCI, mild cognitive impairment; AD, Alzheimer’s disease; FDR, false discovery rate; Αβ, amyloid-β; GFAP, glial fibrillary acidic protein; NfL, neurofilament light; p-tau, phosphorylated tau

Fig. 3

Associations of plasma biomarkers with AD pathologies and discrimination of Aβ status. Associations of plasma biomarkers with AD pathologies were tested using multiple linear regressions adjusted for age and sex. A Associations of plasma levels with CSF AD core biomarkers were tested in CU A−T− and preclinical AD (CU A+T−, CU A+T+) groups (the upper part) and in CU A−T− and prodromal AD (CU A+T−, CU A+T+, and MCI+) groups (the lower part). Colors represent correlation coefficients (β) in regression models and the color bar represents the range of β values obtained. Significance: ***p<0.001, **p<0.01, *p<0.05, −: p≥0.05 (FDR-corrected). B Relationships between plasma levels and cerebral amyloid burden as assessed on Aβ (¹⁸F-AV45) PET images. Aβ PET standard uptake value ratios (SUVRs) of different brain regions were extracted beforehand. Plasma GFAP and p-tau181 showed remarkably positive associations with Aβ PET SUVRs in many brain regions among MCI+ and AD dementia patients. Colors represent t values in regression models and the color bar represents the range of t values obtained. C Receiver operating curve analyses to discriminate Aβ-positive from Aβ-negative status. The upper and bottom plots showed the AUCs of different plasma biomarkers (plasma Aβ42/Aβ40, Aβ42, Aβ40, p-tau181, NfL, and GFAP) in discriminating Aβ status defined by CSF Aβ42 and Aβ PET, respectively. Abbreviations: AD, Alzheimer’s disease; Αβ, amyloid-β; CSF, cerebrospinal fluid; MCI, mild cognitive impairment; FDR, false discovery rate; GFAP, glial fibrillary acidic protein; p-tau, phosphorylated tau; PET, positron emission tomography; CU, cognitively unimpaired; NfL, neurofilament light; AUC, area under the curve

Fig. 4

Trajectories of the biomarkers. Dynamic changes of the biomarkers as a function of CSF Aβ42/Aβ40 (A) or p-tau181 (B) in CU A−T− and preclinical AD stages. The graphs represent the z-score changes of plasma biomarkers using the CU A−T− group as a reference. For comparison purposes, the trajectories of CSF biomarkers were delineated. The solid curves depict biomarker trajectories using robust local weighted regressions. The dashed lines depict the cutoffs for CSF Aβ42/Aβ40 and p-tau. The horizontal axis direction of CSF Aβ42/Aβ40 (A) was inverted. C The relationship between each biomarker and CSF Aβ42/Aβ40 and p-tau181 in CU A−T− and preclinical AD individuals. For each biomarker, we computed the linear regression coefficients (β) and p-values of the z-scores in each of the negative or positive groups. The horizontal grey line represents the absolute value of β, and the color of the solid circle represents the positive (orange) or negative (blue) direction of β. The black circle outside the solid circle indicates the regression p-value<0.05. When the regression slopes between the negative and positive groups were different, a “*” was marked. D Dynamic changes of the biomarkers from CU A−T− to CU A+T*, MCI+, and AD dementia stages. The z-score changes of biomarkers using the CU A−T− group as a reference were delineated. E A model of the approximative orderings of AD-related biomarkers. Approximate stages of the Alzheimer’s continuum (preclinical, prodromal, and AD dementia) were shown on the x-axis. Considering personal reserves and vulnerability factors, we acknowledge that there may be substantial interindividual variations in the timing of different events. Abbreviations: CSF, cerebrospinal fluid; Αβ, amyloid-β; p-tau, phosphorylated tau; AD, Alzheimer’s disease; GFAP, glial fibrillary acidic protein; PET, positron emission tomography; NfL, neurofilament light; MCI, mild cognitive impairment; MRI, magnetic resonance imaging; FDG, fluorodeoxyglucose