Different neuroinflammatory spatial distribution and pathological correlation in clinically defined AD variants

BACKGROUND: Alzheimer's disease (AD) is a clinically and pathologically heterogenous disease. The atypical (non-amnestic) clinical variants are associated with a different distribution of neurofibrillary tangles (NFTs). Neuroinflammation being a key process in AD, is shown to be both correlated to NFTs and Ab plaques. In the current study we assess the spatial neuroinflammatory profile and its pathological hallmark correlate in clinically (a)typical AD. METHOD: AD cases were derived from the Amsterdam Dementia Cohort and categorized as either typical (memory-dominated presentation) or atypical (non-memory presentation) by a cognitive neurologist, resulting in 10 typical and 10 atypical AD cases and cases were compared to 10 non-neurological controls (Fig. 1 for inclusion process; Table 1 for group demographics including clinical and pathological information). Post-mortem brain tissue, including limbic and neocortical regions was collected and processed for immunohistochemistry (IHC) against pTau, Ab, reactive astrocytes (GFAP), and (activated) microglia (Iba1, CD68, MHC-II). Immunostained slides were digitized and quantified for % immunoreactive area. A linear mixed model appropriately corrected was used to test if each IHC marker distribution over the regions within groups was different. A Pearson correlation was performed to test for correlations between pathological hallmarks and neuroinflammation. RESULT: See Fig. 2 for heatmaps of the regional distribution of each IHC marker per diagnostic group. Due to the atypical phenotype variability, the results are also shown for the two major phenotypic groups, being classified as 'frontal' (dysexecutive + behavioral) and 'parietal' (wordfinding + visuospatial). Distribution patterns of all neuroinflammatory markers differed between typical and atypical AD. The most interesting findings were that GFAP immunoreactivity in typical AD correlated with both pTau and Ab, whereas in atypical AD only a correlation with Ab was found. In typical AD, MHC-II correlated negatively with pTau and Ab, whereas in atypical AD MHC-II correlated positively with pTau. CONCLUSION: Not only neuroinflammatory distribution patterns differ between typical and atypical AD, the correlation between neuroinflammation and the specific pathology (pTau or Ab) also differs, suggesting a difference in disease etiology between AD subtypes.