Plaque, Amyloid

Our statistical plan follows. All analyses were conducted in SAS 9.4.

We first conducted a series of Chi square analyses in order to determine if there were disproportionate frequencies of one or another APOE genotype namely “e2” (comprised of e2/e2s and e2/e3 cases), “e3” (e3 homozygotes), “e3/e4” cases, and “e4/e4” cases associated with neuropathological staging or presence/absence of pathology. The e2/e4 genotype (N = 46 cases) was examined in a separate series of analyses.

If findings were positive, we refined our analysis by conducting two planned contrasts in regression models in which e2 was contrasted with e3 and e2 was contrasted with e4 as predictors. In these regressions we adjusted for age at death and sex. If the outcome measure was binary, we utilized logistic regression. If the outcome was ordinal, we utilized ordinal regression.

Given the number of Chi-square analyses that we conducted (12) we used a Bonferroni correction to reduce the probability of type I error. Thus, for 12 Chi square analyses, we set significance at p < 0.004. We considered 0.004 < p < 0.01 trend level significance. For the two planned contrasts using ORs (e2 v e3 and e2 v e4), we considered p < 0.01 as significant. p values for ORs were derived from maximum likelihood estimate Wald Chi squares.
We elected to examine APOE genotype associations with the following classes of neuropathologies.
(1) AD-related pathologies based on the robust association of e2 with reduced risk of clinically diagnosed AD and e4 with increased risk for AD. Histopathologies were as defined in the Montine ABC criteria for severity. (A) Diffuse amyloid plaque (Thal stage) is a measure of spread of plaque and higher scores indicate greater spread of pathology. (B) Braak stage is a measure of progression of NFTs and higher level stages indicate spread of pathology to neocortex. (C) Neuritic plaques are a hallmark feature of AD and may have more specificity to AD than diffuse plaques, with higher levels indicating greater density of pathology.
We also examined the impact of APOE on NFTs rated by Braak stage severity in mediation analyses in which amyloid neuritic plaque served as the mediator in an indirect path, based on consistent evidence that amyloid plaques develop prior to NFTs in AD.
Thus, if APOE genotype effects on Braak stage were significant, we sought to determine if there was a significant mediation effect (i.e., an indirect effect) between APOE genotype (e2 v e3) and Braak stage using the Sobel statistic, which is optimal for identifying mediation effects in large samples, while also examining the direct effect of APOE genotype on Braak stage.
For the e2/e4 genotype, analyzed separately, we sought to determine if e2 might in some way minimize e4 related AD pathology. This genotype thus includes both the protective and risk variant isoforms. We sought to determine if the protective variant can to some degree moderate the effects of e4, or if e4 is toxic and can promote pathology independent of e2.
(2) Lewy body disease due to alpha-synuclein aggregations, as it has recently been proposed that there is increased co-morbidity between AD and Lewy body disease^{19 (link)}. Insofar as APOE e4 is a driver of AD, it may be predicted that it will be associated with LB dementia. Ratings were based on midbrain only, limbic, and neocortical involvement.
(3) FTLD related protein aggregation pathologies including 3R tau Picks disease, other 4R tauopathies and TDP-43 pathology, given suggestions that either AD pathology may promote other protein aggregation neurodegenerative disorders or that protein aggregation disorders share molecular properties that increase risk of co-morbidity. Hence, if APOE e4 promotes a protein aggregation disorder such as AD, it may also promote other such disorders. Similarly, if e2 reduces risk for a protein aggregation disorder such as AD it may reduce risk for other such disorders. All these pathologies were rated as absent/present.

A separate set of mice was used for these experiments. WT and APP/PS1 mice received cGP intranasally for 28 days. They were sacrificed 4-5 h after receiving the drug on the 28^th day. The animals were anaesthetized with halothane and decapitated. The brains were taken out and washed with ice-cold PBS. The two hemispheres of the brain were separated. The left hemisphere of the brain from half of the animals and the right hemisphere of the brain from other half of the animals were taken and fixed overnight in 4% paraformaldehyde. In the next day, the samples were transferred to a 20% sucrose solution in PBS until they sank to the bottom, after which they were transferred to a 30% sucrose solution in PBS until they sank to the bottom.
The samples were sectioned into 40 μm thick sections using cryostat (Leica CM3050 S) and placed in the wells of 24-well cell culture plates containing PBS with 0.01% sodium azide. Every 6^th section of a hemisphere was mounted on a slide, with a total of 5 sections per hemisphere. A single slide contained mounted sections from all the experimental groups.
Thioflavin-S (Sigma-Aldrich; cat. no. T1892) is a benzothiazole dye that displays enhanced fluorescence on binding to fibrillar β-sheets. The sections were stained with thioflavin-S to visualize amyloid plaques [10 (link)]. The stained sections were imaged using a Leica DFC 320 camera connected to a Leica DM RXA2 microscope (excitation at 390 nm and emission at 428 nm). All the imaging parameters were set initially with the first section, and they were kept constant for all the later sections across all the groups. The plaque load was quantified using ImageJ (NIH). Regions of interest were defined using free-hand tool, along the contours of the hippocampus and cortex. The RGB images were converted to binary images (8 bits). Thresholding values were manually adjusted to identify the plaques. To make sure that the specified value incorporates all the plaques, the thresholded image was compared with the original RGB image. Quantification of the plaques was done using “Analyze Particles” plugin of ImageJ, and the following four parameters were evaluated: plaque density (plaque count per mm²), percentage area covered by plaques, size of remaining plaques, and total area defined for analysis. Raw data values from all the five sections of a particular hemisphere were averaged to get a single value for each animal. Only in the representative images shown in the figures, for clarity, the brightness and contrast values were adjusted identically in both groups.

The primary outcome measure is K-MMSE score changes at the endpoint. The secondary outcome measures include demographic data, baseline and follow-up SNSB subdomain scores, self-report questionnaires, HCT scores, brain MRI markers, florbetaben PET positivity, quantitative regional amyloid depositions using PET scans, and clinical progression rates.
All participants will undergo baseline neurologic examinations, neuropsychological tests named SNSB, brain MRIs, blood labs, and florbetaben PET scans for amyloid depositions. PET findings are interpreted using a visual rating scale named brain amyloid plaque load and rated as positive amyloidosis with a brain amyloid plaque load score of 2/3.^{[9 (link)]} Quantitative neuroimaging analysis will be performed.
At baseline, questionnaires for SCD, amyloid PET scans, brain MRIs including 3 dimensional-T1 imaging, plasma amyloid beta values are examined. Telephone-based HCT at home are performed every 6 months during the study period. Annual follow-up evaluations include detailed neuropsychological tests, physical and neurologic examinations, and physician’s assessments for clinical progression. Brain MRI and plasma amyloid beta values are assessed at baseline and 24 months later (Table 1).
Clinical progression to mild cognitive impairment or dementia will be assessed at the final visit. The cognitive tests were administered by a trained neuropsychologist. Participants with CDR score ≥ 0.5 or Korean version of activities of daily living score ≥ 0.43 were considered to have progressed to MCI or dementia.