Idiotypic-susceptible Alzheimer’s disease: a clinically relevant, neurofibrillary tangle subtype

Abstract

Neurofibrillary tangles in Alzheimer’s disease (AD) stereotypically spread from the medial temporal lobe to association areas and then to idiotypic areas (i.e., primary motor, somatosensory, auditory, and visual). Previous studies have reported variable and clinically relevant tangle densities across the hippocampus and association cortices, but the idiotypic tangle burden is understudied. In this study, we measured tangle density using immunohistochemistry in three idiotypic cortices (primary motor, somatosensory, and visual), three association cortices (middle frontal, superior temporal, and inferior parietal), and two hippocampal sectors (CA1 and subiculum) in 144 cases with a high level of AD neuropathologic change. Clinical diagnoses included late-onset AD (LOAD, n = 50), early-onset AD (EOAD, n = 21), behavioral variant frontotemporal dementia (bvFTD, n = 19), corticobasal syndrome (CBS, n = 18), logopenic primary progressive aphasia (lvPPA, n = 21), and posterior cortical atrophy (PCA, n = 15). We algorithmically assigned cases outside the interquartile ranges of mean tangle ratios of association:hippocampal, idiotypic:association, and idiotypic:hippocampal to mutually exclusive subtypes: idiotypic-susceptible, associative-predominant, limbic-predominant, or typical Braak (for all remaining cases). Regional tangle burdens differentiated subtypes, while female sex, younger ages, and longer disease durations also influenced tangle severity. Compared to typical Braak cases, idiotypic-susceptible and associative-predominant cases exhibited shorter disease duration and younger age at death while limbic-predominant cases were older. The MAPT H1H1 haplotype also differed by subtype, being most prevalent in limbic-predominant and least common in idiotypic-susceptible and associative-predominant subtypes. Clinically, the idiotypic-susceptible subtype associated with CBS (56%), the associative-predominant subtype with bvFTD (53%), and the limbic-predominant subtype with LOAD (14%). The typical Braak subtype characterized 74–76% of amnestic AD cases and 32–53% of non-amnestic AD cases. Moreover, k-means clustering corroborated four clusters including the idiotypic-susceptible and associative-predominant patterns. Our results confirm previously described tau subtypes and describe an idiotypic-predominant subtype with clinical relevance and distinct demographic and genetic characteristics in AD.