Ante-mortem clinical history for age at onset of cognitive symptoms, education, and Mini-Mental State Examination (MMSE) score (Folstein et al., 1975 (link)) was retrospectively abstracted from clinical reports for the Mayo Clinic Jacksonville series, with investigators blinded to Thal amyloid phase. Time elapsed from age of onset and last MMSE to death was calculated by subtracting the date of onset or MMSE test date, respectively. Time was converted to years by dividing by 365.25. The final MMSE score was recorded if MMSE test date was within 3 years of death.
Upon neuropathologic examination, Mayo Clinic Jacksonville brains were received for evaluation with the left hemibrain formalin-fixed and the right frozen at −80°. Brain weight represents the fixed specimen that was calculated based upon doubling the weight of the available (usually left) hemibrain. A standardized dissection and sampling method was used as described previously (Terry et al., 1987 (link)). Tissue samples were processed and embedded in paraffin blocks. Mayo Clinic Rochester brains were sampled and examined according to the CERAD protocol (Mirra et al., 1991 (link)). Comparable cortical and subcortical slides were sent to Mayo Clinic Jacksonville for staining and assessment. Senile plaques and NFT were assessed and severity of amyloid angiopathy scored with thioflavin-S fluorescent microscopy, as previously described (Murray et al., 2011 (link)). The thioflavin-S staining protocol we used for these studies is sensitive to all senile plaque types (e.g. diffuse, cored, and neuritic) (Dickson et al., 1992 (link)), which were each included with a truncated maximum of 50 plaques per 3 mm2 using a ×10 objective (Supplementary Fig. 1 ). At the time of diagnosis Mayo Clinic Jacksonville brains were assigned a Braak NFT stage using thioflavin-S (Braak and Braak, 1991 (link)), with retrospective assessment on Mayo Clinic Rochester brains performed subsequent to thioflavin-S staining. Thal amyloid phase was assigned for Mayo Clinic Jacksonville brains by retrospectively assessing senile plaque quantification of neocortex (i.e. mid-frontal, inferior parietal or superior temporal cortex) and hippocampus using thioflavin-S staining results as supported in the latest NIA-AA recommendations (Hyman et al., 2012 (link); Montine et al., 2012 (link)). Neuropathologic reports were abstracted for non-database material on basal ganglia and cerebellum to complete Thal amyloid phase evaluation (Thal et al., 2002 (link)). We do not currently evaluate senile plaques in the superior colliculus or substantia nigra, but the CA4 of the hippocampus performs as well if not better (Thal et al., 2002 (link)). The maximum Thal amyloid phase was assigned if senile plaques were found in: Phase 1: neocortex; Phase 2: CA1/subiculum of the hippocampus; Phase 3: basal ganglia; Phase 4: CA4 of the hippocampus; and Phase 5: cerebellar molecular layer. Supplementary Fig. 1 illustrates regional assessment of senile plaques on thioflavin-S microscopy for Thal amyloid phase. TAR DNA binding protein 43 (TARDBP, previously known as TDP-43) positivity (Amador-Ortiz et al., 2007 (link)) and Lewy body disease type (Uchikado et al., 2006 (link)) was assessed using immunohistochemical methods, as previously described. The presence of significant vascular disease was based on Jellinger and Attems (2003) (link) incidence study. Global cerebral amyloid angiopathy was assessed using a semiquantitative severity measure of 0 = none; 1 = mild; 2 = moderate; 3 = severe from the most severely affected region.
PET imaging was performed in Mayo Clinic Rochester using the 11C amyloid tracer PiB with four 5-min dynamic frames acquired 40–60 min after injection, as previously described (Jack et al., 2008 (link); Kantarci et al., 2012a (link)). Standard corrections, co-registrations, and normalization to internal references were applied (Jack et al., 2008 (link); Kantarci et al., 2012a (link)). Briefly, PiB-PET images were co-registered to the T1-weighted MRI scan of the subject with a custom modified anatomical labelling atlas. Atlas labels in the custom template space were warped to the native T1 MRI space of the subject, as previously described (Kantarci et al., 2012a (link)). PiB-PET cortical regions of interest were partial volume corrected and included both grey matter and white matter without segmentation. PiB-PET uptake was normalized to cerebellar grey/white matter uptake to obtain SUV ratios for brain regions. The meta-region of interest used to measure cortical retention of PiB included an average of the bilateral prefrontal, orbitofrontal, temporal, parietal, anterior cingulate, posterior cingulate, and precuneus regions (Jack et al., 2008 (link)). Using our pipeline we defined the threshold for PiB-positivity at the cut-off point that corresponded to a 90% sensitivity of clinically diagnosed Alzheimer’s disease subjects with an abnormal PET scan (Jack et al., 2014 (link)), which corresponds to a PiB-PET SUV ratio ≥ 1.4.
For APOE genotyping, DNA was obtained from frozen brain tissue using standard protocols. Each sample was genotyped for APOE-ε2, -ε3, and -ε4 with TaqMan® chemistry (Applied Biosystems).
Upon neuropathologic examination, Mayo Clinic Jacksonville brains were received for evaluation with the left hemibrain formalin-fixed and the right frozen at −80°. Brain weight represents the fixed specimen that was calculated based upon doubling the weight of the available (usually left) hemibrain. A standardized dissection and sampling method was used as described previously (Terry et al., 1987 (link)). Tissue samples were processed and embedded in paraffin blocks. Mayo Clinic Rochester brains were sampled and examined according to the CERAD protocol (Mirra et al., 1991 (link)). Comparable cortical and subcortical slides were sent to Mayo Clinic Jacksonville for staining and assessment. Senile plaques and NFT were assessed and severity of amyloid angiopathy scored with thioflavin-S fluorescent microscopy, as previously described (Murray et al., 2011 (link)). The thioflavin-S staining protocol we used for these studies is sensitive to all senile plaque types (e.g. diffuse, cored, and neuritic) (Dickson et al., 1992 (link)), which were each included with a truncated maximum of 50 plaques per 3 mm2 using a ×10 objective (
PET imaging was performed in Mayo Clinic Rochester using the 11C amyloid tracer PiB with four 5-min dynamic frames acquired 40–60 min after injection, as previously described (Jack et al., 2008 (link); Kantarci et al., 2012a (link)). Standard corrections, co-registrations, and normalization to internal references were applied (Jack et al., 2008 (link); Kantarci et al., 2012a (link)). Briefly, PiB-PET images were co-registered to the T1-weighted MRI scan of the subject with a custom modified anatomical labelling atlas. Atlas labels in the custom template space were warped to the native T1 MRI space of the subject, as previously described (Kantarci et al., 2012a (link)). PiB-PET cortical regions of interest were partial volume corrected and included both grey matter and white matter without segmentation. PiB-PET uptake was normalized to cerebellar grey/white matter uptake to obtain SUV ratios for brain regions. The meta-region of interest used to measure cortical retention of PiB included an average of the bilateral prefrontal, orbitofrontal, temporal, parietal, anterior cingulate, posterior cingulate, and precuneus regions (Jack et al., 2008 (link)). Using our pipeline we defined the threshold for PiB-positivity at the cut-off point that corresponded to a 90% sensitivity of clinically diagnosed Alzheimer’s disease subjects with an abnormal PET scan (Jack et al., 2014 (link)), which corresponds to a PiB-PET SUV ratio ≥ 1.4.
For APOE genotyping, DNA was obtained from frozen brain tissue using standard protocols. Each sample was genotyped for APOE-ε2, -ε3, and -ε4 with TaqMan® chemistry (Applied Biosystems).
