Sequence and trajectory of early Alzheimer’s disease-related tau inclusions in the hippocampal formation of cases without amyloid-β deposits
Heiko Braak, Benjamin Mayer, Simone Feldengut, Michael Schön, Kelly Del Tredici
Abstract
Sporadic Alzheimer's disease (AD) involves specific neuronal types and progresses in a systematic manner, permitting subdivision into six neuropathological stages. Neurofibrillary tangle (NFT) stages I-III display abnormal tau inclusions confined to subcortical nuclei and temporal allocortical regions, frequently without amyloid β (Aβ) deposition. We previously suggested a sequence of neuronal involvement in AD that could proceed from entorhinal pre-α cells to hippocampal prosubicular pyramidal cells and the CA1/CA2 sectors, from there to the thorny excrescences on mossy cells in CA3/CA4, and, finally, from the mossy cells to dentate fascia (Fd) granular cells. Here, we aimed to see if associations existed between the early NFT stages I-III, when Aβ deposits are frequently absent, and the following four categories: (1) anatomical regions and abnormal morphological tau changes in region-specific layers, (2) nerve cell loss, (3) APOE genotype, and (4) the trajectory (directionality) of tau progression in the hippocampal formation. To do so, we examined the transentorhinal/entorhinal regions and hippocampal formation using AT8-immunohistochemistry in 100 µm sections from N = 308 brains with tau inclusions lacking Aβ deposits between NFT stages I and III (average age at death 66.7 years for females, 66.4 years for males). Our results indicated a significantly (p < 0.001) ordered progression of abnormal tau in a direction opposite to currently known unidirectional intrahippocampal connections, thereby indirectly supporting the idea of transneuronal abnormal tau spreading, i.e., anterogradely, through the hippocampal formation. Tau-related neuronal loss was also significant (p < 0.001 for the transentorhinal/entorhinal regions and for sectors CA1/CA2 and p = 0.003 for CA3/CA4/Fd). These findings challenge the amyloid cascade and the PART hypotheses, corroborating the concept that early AD-related tau inclusions and tau-related neuronal loss occur independently of Aβ deposition.