Developing Topics.

Abstract

Limbic-predominant age-related TDP-43 encephalopathy neuropathological change (LATE-NC) is found in over half of pathologically confirmed Alzheimer's disease (AD) dementia cases. However, the region-specific relationship between AD neuropathologic change (ADNC) and LATE-NC remains unclear. Here, we investigated the intraregional and interregional associations between ADNC and LATE-NC. We analyzed multi-regional postmortem neuropathology data from the Religious Orders Study and the Rush Memory and Aging Project (ROSMAP). Quantitative ADNC burden was assessed with amyloid-β (Aβ) and tau immunohistochemistry in 8 brain regions. The Aβ burden was averaged across all regions, given strong inter-regional correlations. Semi-quantitative LATE-NC burden was assessed with TDP-43 immunohistochemistry in 7 brain regions, scored in 6 grades. Hippocampal sclerosis (HS) was defined as a severe neuronal loss in a mid-hippocampal coronal section, irrespective of ADNC or LATE-NC burden. We assessed the association between regional LATE-NC and global Aβ burden, adjusting for APOE ε4, age at death, and sex. Then, we tested the intra- and inter-regional associations between LATE-NC and tau, adjusting for Aβ, APOE ε4, age at death, and sex. We performed a post-hoc subgroup analysis in participants without HS. We analyzed the data from n = 1264 deceased ROSMAP participants (Table 1). Global Aβ was positively associated with LATE-NC in most regions (Fig.1A). Interestingly, even after accounting for Aβ and APOE ε4, tau and LATE-NC were positively associated within the entorhinal (t = 7.29, p = 5.4×10-13) and midfrontal cortex (t = 4.33, p = 1.6×10-5) but not within the hippocampus (Fig.1B, bold boxes). Tau and LATE-NC were positively correlated across most interregional pairs (Fig.1B) except for negative associations between hippocampal tau (CA1/subiculum) and frontal LATE-NC. There was no association between hippocampal tau and frontal LATE-NC in the HS(-) subgroup (n = 1143; inferior frontal, p = 0.67; midfrontal, p = 0.15), suggesting that extensive hippocampal neuronal loss in HS(+) cases could have artifactually driven the negative association between hippocampal tau and frontal LATE-NC. Aβ was positively associated with LATE-NC in most regions. Even after adjusting for Aβ and APOE ε4, tau was positively associated with LATE-NC across most regions, except for hippocampal tau and frontal LATE-NC. Further studies are required to determine the underlying mechanisms linking ADNC and LATE-NC.