Cristian E Leyton1, Ben Cassidy2, Victor L Villemagne3, Gareth Jones4, John B Kwok5, Christopher C Rowe4, Kirrie J Ballard6, Olivier Piguet7, John R Hodges7. 1. Faculty of Health Sciences, The University of Sydney, Lidcombe; Neuroscience Research Australia Randwick; Australian Research Council Centre of Excellence in Cognition and its Disorders, Sydney, New South Wales, Australia. Electronic address: cristian.leyton@sydney.edu.au. 2. Department of Biostatistics, Mailman School of Public Health, Columbia University, New York, New York. 3. Department of Nuclear Medicine and Centre for PET, Austin Health, Heidelberg; The Florey Institute of Neuroscience and Mental Health, The University of Melbourne, Melbourne, Victoria. 4. Department of Nuclear Medicine and Centre for PET, Austin Health, Heidelberg. 5. Neuroscience Research Australia Randwick. 6. Faculty of Health Sciences, The University of Sydney, Lidcombe. 7. Neuroscience Research Australia Randwick; School of Medical Sciences, The University of New South Wales, Sydney, New South Wales, Australia.
Abstract
BACKGROUND: Despite divergent clinical features, language and amnestic presentations of Alzheimer's disease (AD) appear to show comparable regional amyloid-β (Aβ) burden. By using a statistical network approach, we aimed to identify complex network patterns of Aβ deposition and explore the effect of apolipoprotein E (APOE) ε4 allele on cortical Aβ burden across AD phenotypes. METHODS: Sixteen amnestic AD participants and 18 cases with logopenic-variant of primary progressive aphasia (lv-PPA) with a high cortical Aβ burden were selected. A comprehensive clinical assessment, Aβ imaging, and APOE genotyping were performed in all cases. Statistical network analysis was undertaken based on the estimation of sparse partial correlations of Aβ burden between cortical regions. Global and regional network statistical parameters as well as the effect of APOEε4 genotype on cortical Aβ were explored. RESULTS: The two groups showed equivalent distribution of cortical amyloid burden and frequency of APOEε4 genotype. Statistical network analysis, however, demonstrated divergent connectivity properties. The lv-PPA group demonstrated higher mean network degree and shorter characteristic path length than the amnestic AD group. Amnestic AD cases showed connectivity hubs confined to the mesial temporal and prefrontal lobes bilaterally, whereas lv-PPA cases showed hubs scattered across the whole cortical mantle. An interaction effect on total Aβ burden between APOE genotype and AD presentations was also detected. CONCLUSIONS: The network analysis reveals interregional network differences not evident using a simple comparison of Aβ burden. This suggests that regional neurotoxic effects may explain the phenotypical differences in AD presentation and that these can be modulated by APOE genotype.
BACKGROUND: Despite divergent clinical features, language and amnestic presentations of Alzheimer's disease (AD) appear to show comparable regional amyloid-β (Aβ) burden. By using a statistical network approach, we aimed to identify complex network patterns of Aβ deposition and explore the effect of apolipoprotein E (APOE) ε4 allele on cortical Aβ burden across AD phenotypes. METHODS: Sixteen amnestic ADparticipants and 18 cases with logopenic-variant of primary progressive aphasia (lv-PPA) with a high cortical Aβ burden were selected. A comprehensive clinical assessment, Aβ imaging, and APOE genotyping were performed in all cases. Statistical network analysis was undertaken based on the estimation of sparse partial correlations of Aβ burden between cortical regions. Global and regional network statistical parameters as well as the effect of APOEε4 genotype on cortical Aβ were explored. RESULTS: The two groups showed equivalent distribution of cortical amyloid burden and frequency of APOEε4 genotype. Statistical network analysis, however, demonstrated divergent connectivity properties. The lv-PPA group demonstrated higher mean network degree and shorter characteristic path length than the amnestic AD group. Amnestic AD cases showed connectivity hubs confined to the mesial temporal and prefrontal lobes bilaterally, whereas lv-PPA cases showed hubs scattered across the whole cortical mantle. An interaction effect on total Aβ burden between APOE genotype and AD presentations was also detected. CONCLUSIONS: The network analysis reveals interregional network differences not evident using a simple comparison of Aβ burden. This suggests that regional neurotoxic effects may explain the phenotypical differences in AD presentation and that these can be modulated by APOE genotype.