Neuropathologic, biochemical, and molecular characterization of the frontotemporal dementias.

The frontotemporal dementias (FTDs) are a heterogeneous group of neurodegenerative disorders that are characterized clinically by dementia, personality changes, language impairment, and occasionally extrapyramidal movement disorders. Historically, the diagnosis and classification of FTDs has been fraught with difficulties, especially with regard to establishing a consensus on the neuropathologic diagnosis. Recently, an international group of scientists participated in a consensus conference to develop such neuropathologic criteria. They recommended a diagnostic classification scheme that incorporated a biochemical analysis of the insoluble tau isoform composition, as well as ubiquitin immunohistochemistry. The use and reliability of this classification system has yet to be examined. In this study, we evaluated 21 cases of FTD. Using traditional histochemical stains and tau protein and ubiquitin immunohistochemistry, we separated each case into one of the following categories: classic Pick disease (PiD; n = 7), corticobasal degeneration (CBD; n = 5), dementia lacking distinctive histopathologic features (DLDH; n = 4), progressive supranuclear palsy (PSP; n = 2), frontotemporal lobar degeneration with motor neuron disease or motor neuron disease-type inclusions (FTLD-MND/MNI; n = 2), and neurofibrillary tangle dementia (NFTD; n = 1). Additionally, we independently categorized each case by the insoluble tau isoform pattern, including 3R (n = 5), 4R (n = 7), 3R/4R (n = 3), and no insoluble tau (n = 6). As suggested by the proposed diagnostic scheme, we found that the insoluble tau isoform patterns correlated strongly with the independently derived histopathologic diagnoses (p < 0.001). The data show that cases containing predominantly 3R tau were classic PiD (100%). Cases with predominantly 4R tau were either CBD (71%) or PSP (29%). Cases with both 3R and 4R tau were either a combination of PiD and Alzheimer disease (67%) or NFTD (33%). Finally, cases with no insoluble tau were either DLDH (67%) or FTLD-MND/MNI (33%). To further characterize these cases, we also performed quantitative Western blots for soluble tau, APOE genotyping, and, in selected cases, tau gene sequencing. We show that soluble tau is reduced in DLDH and FTLD-MND/MNI and that APOE4 is overrepresented in PiD and DLDH. We also identified a new family with the R406W mutation and pathology consistent with NFTD. This study validates the recently proposed diagnostic criteria and forms a framework for further refinement of this classification scheme.