Amygdala atrophies in specific subnuclei in preclinical Alzheimer's disease

Abstract

Introduction: Magnetic resonance imaging (MRI) segmentation algorithms make it possible to study detailed medial temporal lobe (MTL) substructures as hippocampal subfields and amygdala subnuclei, offering opportunities to develop biomarkers for preclinical Alzheimer's disease (AD).

Methods: We identified the MTL substructures significantly associated with tau-positron emission tomography (PET) signal in 581 non-demented individuals from the Alzheimer's Disease Neuroimaging Initiative (ADNI-3). We confirmed our results in our UCLouvain cohort including 110 non-demented individuals by comparing volumes between individuals with different visual Braak's stages and clinical diagnosis.

Results: Four amygdala subnuclei (cortical, central, medial, and accessory basal) were associated with tau in amyloid beta-positive (Aβ+) clinically normal (CN) individuals, while the global amygdala and hippocampal volumes were not. Using UCLouvain data, we observed that both Braak I-II and Aβ+ CN individuals had smaller volumes in these subnuclei, while no significant difference was observed in the global structure volumes or other subfields.

Conclusion: Measuring specific amygdala subnuclei, early atrophy may serve as a marker of temporal tauopathy in preclinical AD, identifying individuals at risk of progression.

Highlights: Amygdala atrophy is not homogeneous in preclinical Alzheimer's disease (AD). Tau pathology is associated with atrophy of specific amygdala subnuclei, specifically, the central, medial, cortical, and accessory basal subnuclei. Hippocampal and amygdala volume is not associated with tau in preclinical AD. Hippocampus and CA1-3 volume is reduced in preclinical AD, regardless of tau.

Keywords: AV1451; Alzheimer's disease; MK6240; amygdala subnuclei; biomarker; florquinitau; flortaucipir; preclinical AD; structural magnetic resonance imaging; tauopathy; tau‐positron emission tomography.

FIGURE 1

Amygdala subnuclei and hippocampus subfield segmentation. Amygdala (left, dotted) and hippocampus (right, solid) substructure segmentation from FreeSurfer version 7.2.

FIGURE 2

Association between volumes and temporal tau in ADNI cohort. Partial Spearman correlation analysis illustrating association between volume of medial temporal lobe structures and temporal lobe tau‐PET AV1451 SUVR in ADNI cohort. The regions considered include the amygdala, the hippocampus, and the substructures for which the R ² was greater than 0.01 (two‐tail p value < 0.20) in cognitively normal Aβ+ subjects, that is, the central, cortical, medial, and accessory basal nuclei of the amygdala. An AA resulting from the combination of these tau‐associated amygdala subnuclei is also represented. Correlations are represented using the bioclinical classification. AA, amygdala aggregate; ADNI, Alzheimer's Disease Neuroimaging Initiative; SUVR, standardized uptake value ratio.

FIGURE 3

Association between volumes and temporal tau in UCLouvain cohort. Partial Spearman correlation analysis illustrating association between structure volume and temporal lobe tau‐PET MK6240 SUVR in UCLouvain cohort. Structures examined include the amygdala, hippocampus, four amygdala subnuclei (cortical, central, medial, and accessory‐basal), and the previously defined amygdala aggregate. Individual data points are discriminated by bioclinical status (color) and visual Braak stage (shape) The correlation coefficient was also adjusted for the volume of the global structure (adjusted R ²). SUVR, standardized uptake value ratio.

FIGURE 4

Volume difference according to visual Braak stages. Boxplots illustrate volume differences according to visual Braak stages in subjects from UCLouvain cohort. The structures examined include the amygdala, the hippocampus, four amygdala subnuclei (cortical, central, medial, and accessory‐basal nuclei), and the previously defined amygdala aggregate (AA). Individual data points are differentiated according to bioclinical status (color) and Braak visual stage (shape). Reported p values are p values adjusted for age, sex, intracranial volume, and years of education: ^* p < 0.05; ^** p < 0.01; ^*** p < 0.001; ****p < 0.0001. [Correction added on October 5, 2024, after first online publication: Figure 4 has been replaced.]

FIGURE 5

Volume difference according to amyloid and clinical classification (global structures and amygdala subnuclei). Boxplots illustrate volume differences according to bioclinical status (including Aβ and neuropsychological status) in subjects from UCLouvain cohort. The structures examined include the amygdala, the hippocampus, four amygdala subnuclei (cortical, central, medial, and accessory‐basal nuclei), and the previously defined AA. Individual data points are differentiated according to bioclinical status (color) and Braak visual stage (shape). Reported p values are p values adjusted for age, sex, intracranial volume, and years of education: ^* p < 0.05; ^** p < 0.01; ^*** p < 0.001; ^**** p < 0.0001. AA, amygdala aggregate.

FIGURE 6

Volume difference according to amyloid and clinical classification (hippocampal subfields). Boxplots illustrate volume differences according to bioclinical status (including Aβ and neuropsychological status) in subjects from UCLouvain cohort. The structures examined include hippocampal subfields as CA1, CA3, the DG, and the HATA. Individual data points are differentiated according to bioclinical status (color) and Braak visual stage (shape). Reported p values are p values adjusted for age, sex, intracranial volume, and years of education: ^* p < 0.05; ^** p < 0.01; ^*** p < 0.001; ^**** p < 0.0001. DG, dentate gryrus; HATA, hippocampal amygdaloid transition area.