Neuropathologic Features of Antemortem Atrophy-Based Subtypes of Alzheimer Disease

Abstract

Background and objectives: To investigate whether antemortem MRI-based atrophy subtypes of Alzheimer disease (AD) differ in neuropathologic features and comorbid non-AD pathologies at postmortem.

Methods: From the Alzheimer's Disease Neuroimaging Initiative cohort, we included individuals with antemortem MRI evaluating brain atrophy within 2 years before death, antemortem diagnosis of AD dementia/mild cognitive impairment, and postmortem-confirmed AD neuropathologic change. Antemortem atrophy subtypes were modeled as continuous phenomena based on a recent conceptual framework: typicality (spanning limbic-predominant AD to hippocampal-sparing AD) and severity (spanning typical AD to minimal atrophy AD). Postmortem neuropathologic evaluation included AD hallmarks, β-amyloid, and tau as well as non-AD pathologies, alpha-synuclein and TAR DNA-binding protein 43 (TDP-43). We also investigated the overall concomitance across these pathologies. Partial correlations assessed the associations between antemortem atrophy subtypes and postmortem neuropathologic outcomes.

Results: In 31 individuals (26 AD dementia/5 mild cognitive impairment, mean age = 80 years, 26% females), antemortem typicality was significantly negatively associated with neuropathologic features, including β-amyloid (rho = -0.39 overall), tau (rho = -0.38 regionally), alpha-synuclein (rho = -0.39 regionally), TDP-43 (rho = -0.49 overall), and concomitance of pathologies (rho = -0.59 regionally). Limbic-predominant AD was associated with higher Thal phase, neuritic plaque density, and presence of TDP-43 compared with hippocampal-sparing AD. Regionally, limbic-predominant AD showed a higher presence of tau and alpha-synuclein pathologies in medial temporal structures, a higher presence of TDP-43, and concomitance of pathologies subcortically/cortically compared with hippocampal-sparing AD. Antemortem severity was significantly negatively associated with concomitance of pathologies (rho = -0.43 regionally), such that typical AD showed higher concomitance of pathologies than minimal atrophy AD.

Discussion: We provide a direct antemortem-to-postmortem validation, highlighting the importance of understanding atrophy-based heterogeneity in AD relative to AD and non-AD pathologies. We suggest that (1) typicality and severity in atrophy reflect differential aspects of susceptibility of the brain to AD and non-AD pathologies; and (2) limbic-predominant AD and typical AD subtypes share similar biological pathways, making them more vulnerable to AD and non-AD pathologies compared with hippocampal-sparing AD, which may follow a different biological pathway. Our findings provide a deeper understanding of associations of atrophy subtypes in AD with different pathologies, enhancing the prevailing knowledge of biological heterogeneity in AD and could contribute toward tracking disease progression and designing clinical trials in the future.

Figure 1. Distribution of (A) Antemortem MRI-Based Heterogeneity and (B-C) Postmortem Neuropathology Superposed on MRI-Based Heterogeneity

(A) Antemortem atrophy subtypes modeled as continuous phenomena by the dimensions of typicality and severity. Four individual cases are highlighted, showing the extremes on each dimension; (B) postmortem AD neuropathologic change; and (C) postmortem secondary diagnosis assigned per individual. All plots show antemortem MRI-based typicality on the horizontal scale, proxied by the index = ; all plots show antemortem MRI-based severity on the vertical scale, proxied by the Global Brain Atrophy Index = , whereby higher values correspond to lower severity. AD = Alzheimer disease; ADNC = AD neuropathologic change; ALB = amygdala Lewy bodies; aMCI = amnestic mild cognitive impairment; DLB = dementia with Lewy bodies; HS = hippocampal-sparing AD; HSc = hippocampal sclerosis; ICH = intracerebral hemorrhage; LP = limbic-predominant AD; MA = minimal atrophy AD; RID = assigned individual ID in the AD Neuroimaging Initiative dataset; SAL = subcortical arteriosclerotic leukoencephalopathy; SDH = subdural hemorrhage; TAD = typical AD; TDP-MTL = TAR DNA-binding protein in the medial temporal lobe.

Figure 2. Distribution of Postmortem AD Neuropathologies Superposed on MRI-Based Heterogeneity

Postmortem AD pathologies used to assess ADNC, encompassing the “ABC” scores of (A) Thal phase for Aβ, (B) Braak stage for tau, and (C) Consortium to Establish a Registry for AD neuritic plaques. All plots show antemortem MRI-based typicality on the horizontal scale, proxied by the index = ; all plots show antemortem MRI-based severity on the vertical scale, proxied by the Global Brain Atrophy Index = , whereby higher values correspond to lower severity. AD = Alzheimer disease; ADNC = AD neuropathologic change.

Figure 3. Distribution of Postmortem Non-AD Neuropathologies Superposed on MRI-Based Heterogeneity

Postmortem non-AD pathologies including (A) α-synuclein Lewy bodies and (B) TDP-43. All plots show antemortem MRI-based typicality on the horizontal scale, proxied by the index = ; all plots show antemortem MRI-based severity on the vertical scale, proxied by the Global Brain Atrophy Index = , whereby higher values correspond to lower severity. A + E = TDP-43 immunoreactive inclusions are present in the amygdala and entorhinal/inferior temporal cortex; A + H + E + N = TDP-43 immunoreactive inclusions are present in the amygdala, hippocampus, entorhinal/inferior temporal cortex, and neocortex; A + H + E = TDP-43 immunoreactive inclusions are present in the amygdala, hippocampus, and entorhinal/inferior temporal cortex; AD = Alzheimer disease; TDP-43 = TAR DNA-binding protein 43.

Figure 4. Association Between Antemortem MRI-Based (A) Typicality and (B) Severity and Regional Neuropathologic Features

Associations between each of typicality or severity and presence of regional pathologies were evaluated using linear partial correlation, adjusted for field strength, age at scan, and the other dimension (severity or typicality); linear partial correlation coefficient (rho) and significant p values are indicated. AMYG = amygdala; Aβ = β-amyloid (diffuse and cored plaques); CA1 = hippocampus at the level of lateral geniculate nucleus including cornu Ammonis 1 subfield; DG = hippocampus at the level of lateral geniculate nucleus including dentate gyrus; ERC = entorhinal cortex; IPL = inferior parietal lobe (angular gyrus); MFG = middle frontal gyrus; PHG = hippocampus at the level of lateral geniculate nucleus including parahippocampal gyrus; STG = superior and middle temporal gyri; Tau = phosphorylated tau assessing neurofibrillary tangles; TDP-43 = phosphorylated TAR DNA-binding protein 43 neuronal cytoplasmic inclusion; α-syn = alpha-synuclein Lewy body pathology.

Figure 5. Susceptibility of Antemortem MRI-Based Heterogeneity to AD and Non-AD Neuropathologies

Associations of antemortem typicality and severity with postmortem neuropathologic features may generate the following hypotheses: (1) the orthogonal dimensions of biological heterogeneity, typicality and severity, may offer complementary information regarding the vulnerability of the brain to AD (amyloid, tau) and non-AD (α-syn, TDP-43) pathologies; and (2) limbic-predominant AD along the typicality dimension and typical AD along the severity dimension may share similar underlying biological pathway(s), which make them more susceptible to pathologies, whereas hippocampal-sparing AD along the typicality dimension and minimal AD along the severity dimension may share similar pathway(s), making them less susceptible. α-syn = α-synuclein; AD = Alzheimer disease; TDP-43 = TAR DNA-binding protein 43.