The heterogeneity of asymmetric tau distribution is associated with an early age at onset and poor prognosis in Alzheimer's disease

Abstract

Purpose: Left-right asymmetry, an important feature of brain development, has been implicated in neurodegenerative diseases, although it's less discussed in typical Alzheimer's disease (AD). We sought to investigate whether asymmetric tau deposition plays a potential role in AD heterogeneity.

Methods: Two independent cohorts consisting of patients with mild cognitive impairment due to AD and AD dementia with tau PET imaging were enrolled [the Alzheimer's Disease Neuroimaging Initiative (ADNI) cohort with ¹⁸F-Flortaucipir, the Shanghai Memory Study (SMS) cohort with ¹⁸F-Florzolotau]. Based on the absolute global tau interhemispheric differences, each cohort was divided into two groups (asymmetric versus symmetric tau distribution). The two groups were cross-sectionally compared in terms of demographic, cognitive characteristics, and pathological burden. The cognitive decline trajectories were analyzed longitudinally.

Results: Fourteen (23.3%) and 42 (48.3%) patients in the ADNI and SMS cohorts showed an asymmetric tau distribution, respectively. An asymmetric tau distribution was associated with an earlier age at disease onset (proportion of early-onset AD: ADNI/SMS/combined cohorts, p = 0.093/0.026/0.001) and more severe pathological burden (i.e., global tau burden: ADNI/SMS cohorts, p < 0.001/= 0.007). And patients with an asymmetric tau distribution were characterized by a steeper cognitive decline longitudinally (i.e., the annual decline of Mini-Mental Status Examination score: ADNI/SMS/combined cohorts, p = 0.053 / 0.035 / < 0.001).

Conclusions: Asymmetry in tau deposition, which may be associated with an earlier age at onset, more severe pathological burden, and a steeper cognitive decline, is potentially an important characteristic of AD heterogeneity.

Keywords: Age; Alzheimer’s disease; Asymmetry; Positron emission tomography; Prognosis; Tau.

Fig. 1

Differences in “A/T” burden in patients with asymmetric versus symmetric tau distribution on PET images. (A) Average tau SUVR PET images in patients with asymmetric versus symmetric tau distribution in the ADNI cohort. In presence of an asymmetric tau distribution, images of patients with a left asymmetry were flipped on the X-axis, whereas images of patients with a symmetric tau distribution were randomly flipped on the X-axis. (B-C) In the ADNI cohort, the global tau and amyloid burden was investigated using ¹⁸F-Flortaucipir PET and ¹⁸F-Florbetapir PET imaging, respectively. Three patients whose amyloid PET imaging was performed not during the same period of their tau PET scan were excluded from the comparison of amyloid burden. (D) Average tau SUVR PET images in patients with asymmetric versus symmetric tau distribution in the SMS cohort. The methodology was identical to that implemented in the ADNI cohort. (E-F) In the SMS cohort, the global tau and amyloid burden was investigated using ¹⁸F-Florzolotau PET and ¹⁸F-Florbetapir PET imaging. Four patients with ¹¹C-PIB amyloid PET imaging were excluded from comparisons of the amyloid burden. *, p < 0.05; **, p < 0.01; ***, p < 0.001; n.s., not significant; Mann-Whitney U test. The error bars represent the median (interquartile range). Abbreviations: ADNI, Alzheimer’s Disease Neuroimaging Initiative; SMS, Shanghai Memory Study; SUVR, standardized uptake value ratio; MCI, mild cognitive impairment; AD, Alzheimer’s disease.

Fig. 2

Raw scores and estimated annual changes on neuropsychological testing over time in patients with asymmetric versus symmetric tau distribution. The compound figures consisted of raw spaghetti plots with separate lines showing the unadjusted mean trajectory (with its 95% confidence interval) and the estimated annual changes after adjusting for age at baseline, education (years), APOE ε4 status, and sex for each cohort (A-B). Different cohort was further adjusted when analyzing the combined cohort (C). The spaghetti plots depict the raw scores on neuropsychological tests administered at each visit; the separate lines denote the unadjusted mean trajectory (with its 95% confidence interval) of patients with asymmetric versus symmetric tau distribution. In the ADNI cohort, limited data were available for the last three visits (asymmetric tau distribution: n = 2, 0, 0; symmetric tau distribution: n = 10, 4, 2). Similarly, data concerning the last two visits were limited in the SMS cohort (asymmetric tau distribution: n = 0, 0; symmetric tau distribution: n = 1, 2). Therefore, lines were generated after their exclusion. The bar charts show the estimated annual changes of scores on neuropsychological tests. The p values are calculated for the differences in the slope of longitudinal decline between patients with asymmetric versus symmetric tau distribution using a linear mixed-effect model adjusted for age at baseline, education (years), APOE ε4 status, and sex for each cohort; disease duration was considered as a time scale. Different cohort was further adjusted when analyzing the combined cohort. *, p < 0.05; **, p < 0.01; ***, p < 0.001. Abbreviations: ADNI, Alzheimer’s Disease Neuroimaging Initiative; SMS, Shanghai Memory Study; MMSE, Mini-Mental Status Examination; MOCA, Montreal Cognitive Assessment; CDRSB, Clinical Dementia Rating, Sum of Boxes; FAQ, Functional Assessment Questionnaire; ADAS11, Alzheimer’s Disease Assessment Scale (11 items); BNT, Boston Naming Test.

Fig. 3

Plasma biochemical markers of the “A/T/N” classification system in patients with asymmetric versus symmetric tau distribution in the SMS (¹⁸F-Florzolotau) cohort. **, p < 0.01; ***, p < 0.001; n.s., not significant; Student’s t-test for plasma Aβ42/ Aβ40 ratio, and Mann-Whitney U test for others. The error bars represent the mean (standard deviation) for plasma Aβ42/ Aβ40 ratio, and median (interquartile range) for others. Abbreviations: MCI, mild cognitive impairment; AD, Alzheimer’s disease; Aβ, amyloid-beta protein; P-tau181, tau phosphorylated at threonine 181; T-tau, total tau; NfL, neurofilament protein light chain.