Citations

Abstract Background Multi‐site aging and dementia studies are challenged by data heterogeneity caused by scanner (manufacturer and protocol) related variability. Quantitative analysis of white matter hyperintensity (WMH) from pooled samples, a traditional marker of cerebrovascular disease, is sensitive to machine variability. Our goal was to compare Deming (symmetric) linear regression and ComBat, a common method used for harmonizing neuroimaging measurements, for harmonizing WMH measurements on participants scanned on two different scanners (GE and Siemens) and protocols (2D FLAIR on GE and 3D FLAIR on Siemens). Methods Cognitively unimpaired (CU) ( n =53, age (mean (SD))=67 (15) years) and cognitively impaired (CI; MCI/dementia) ( n =34, age (mean (SD))=82 (9) years) participants from the Mayo Clinic Study of Aging and Mayo ADRC were scanned on GE and Siemens scanners within a week of each other (typically done on the same day). WMH as a percentage of total intracranial volume was measured using multispectral segmentation of FLAIR and T1‐weighted images (Gunter et. al. AAIC 2023). We used Deming regression as it includes errors in both input dimensions. ComBat was applied using age, sex, and diagnosis as covariates, with scanner vendor and protocol as the batch variable. We assessed the inter‐scanner similarity using Maximum Mean Discrepancy (MMD) analysis, where a value of zero indicates identical distributions, p >0.05 indicates no significant difference under H 0 . (Gretton et al., JMLR 2012). Results Our results showed that the distribution of WMH measurements, which were different in the CU group, were more similar after Deming than ComBat harmonization (Figure 1). This was confirmed by the lower MMD in the CU group from Deming regression (‐0.01 ( p =0.76) from 0.03 ( p =0.05)) compared to ComBat (0.04 ( p =0.02))(Figure 1). However, ComBat and Deming were unable to create distributions as compatible as original measurements for the CI group. Conclusion Using a unique dataset of the same participants scanned close in time on different scanner and protocols, we found that (1) MMD measurements were useful to test the need/improvement due to harmonization; (2) ComBat was not substantially better at creating compatible distributions between scanner vendors than Deming regression, and both were sometimes inadequate. Future work will require alternative models to linear scale and shift transformations.

Abstract Background Exposure to repetitive head impacts (RHI) is associated with developing chronic traumatic encephalopathy (CTE) neuropathology. Cognitive and behavioral symptoms have been associated with CTE neuropathology, but efforts to define the specific clinical syndrome are ongoing. This study characterizes the current use of centrally acting medications (CAMs), chronic pain, and number of orthopedic surgeries in former American football players and evaluates for associations with cognitive and behavioral symptoms. Methods This study analyzed data from the DIAGNOSE CTE Research Project, which includes 120 former professional football players, 60 collegiate football players, and 56 asymptomatic men without history of RHI. Medical history was obtained by self‐report, and participants completed the Montreal Cognitive Assessment (MoCA), Brief Pain Inventory, Barratt Impulsiveness Scale‐11 (BIS‐11), BRIEF‐A Behavioral Regulation Index (BRI), Beck Depression Inventory‐II (BDI‐II), Beck Anxiety Index (BAI), and Brown‐Goodwin Lifetime History of Aggression adulthood score (BGLHA). Logistic regression assessed the association of CAMs, average pain score, and orthopedic surgeries on the cognitive impairment and neurobehavioral dysregulation components of the NINDS traumatic encephalopathy syndrome (TES) research criteria, in former American football players. Linear regression assessed the associations with MoCA and behavioral scales. Models were adjusted for age, education, race, and years of football played. Results Tables 1 and 2 describe the sample and CAM burden. More orthopedic surgeries, CAMs, and higher average pain were observed in former American football players. Number of CAMs was associated with neurobehavioral dysregulation (OR = 2.18), lower MoCA (β = ‐0.47), and higher BIS‐11 (β = 2.23), BDI‐II (β = 2.01), BRI (β = 3.71), BAI (β = 2.04), and BGLHA (β = 0.68) scores. Higher average pain was associated with neurobehavioral dysregulation (OR = 1.57) and higher BIS‐11 (β = 2.20), BDI‐II (β = 1.98), BRI (β = 2.20), BAI (β = 1.84), and BGLHA (β = 0.52) scores. No associations were observed with orthopedic surgeries. Conclusions In former American football players, the number of CAMs and higher average pain were associated with neurobehavioral dysregulation. No associations were observed with diagnosis of cognitive impairment by TES criteria, though the number of CAMs was associated with decreased MoCA.

Abstract Background G 4 C 2 repeat expansions in C9orf72 gene are the most common genetic cause of frontotemporal dementia (FTD) and amyotrophic lateral sclerosis (ALS). Mounting evidence indicates that dipeptide repeat (DPR) proteins translated from the expanded repeats, in particular poly(GR) and poly(PR), play a significant role. We and others have shown that these positively‐charged arginine‐rich DPR proteins induce the assembly of stress granules (SGs) and impair their disassembly. While distinct from SGs, processing bodies (PBs) are also cytoplasmic messenger ribonucleoprotein condensates that assemble upon stress. However, the impacts of arginine‐rich DPR proteins on PB assembly and dynamics have yet to be investigated. Method We conducted immunofluorescence staining for arginine‐rich DPR proteins, SG‐ and PB‐associated markers in cultured cells and mouse brain. Additionally, we performed in vitro liquid‐liquid phase separation (LLPS) using recombinant proteins and assessed molecular dynamics of SG and PB through fluorescence recovery (FRAP) after photobleaching in living cells. Result Here, we observed that both GFP‐(GR) 100 and GFP‐(PR) 100 formed cytosolic inclusions immunopositive PB proteins. Unlike GFP‐(GR) 100 and GFP‐(PR) 100 inclusions, cellular cytoplasmic GFP‐(GA) 100 inclusions did not contain PB proteins. We subsequently examined whether the phenomena observed in cultured cells occur in vivo. We observed that PB proteins were recruited to poly(GR) inclusions in cells with aggregated poly(GR). Next, we examined the dynamics of poly(GR)‐induced PBs in cultured cells using the FRAP technique. We observed that the recovery of stress‐induced PBs was more rapid than that of PBs in HEK293T cells expressing GFP‐(GR) 100 . Lastly, mechanistic studies revealed that poly(GR) interacts with PB proteins to form condensates in vitro via LLPS. Conclusion Our results provide both in vitro and in vivo evidence that arginine‐rich DPRs induce the spontaneous formation of PBs via LLPS. Our results also indicate that poly(GR)‐induced PBs are more stable than the transient and dynamic conventional PBs. Given that PBs uniquely enriched with mRNA degradation and decay factors, the consequences of impairing these processes are expected to contribute to abnormal RNA metabolism in c9FTD/ALS.

Abstract Background In vivo Braak staging stratifies patients across the AD spectrum and has the potential to harmonize tau PET tracer staging. This study aims to compare and test harmonization procedures for Braak staging individuals using MK6240 and Flortaucipir tau PET tracers. Methods We assessed 437 participants across the AD spectrum (245 cognitively unimpaired (CU) and 192 cognitively impaired; mean age 68.5 ± 8.6) using head‐to‐head MK6240 and Flortaucipir scans. We computed SUVRs in Braak regions of interest (ROIs) and assessed four cut‐off methods for Braak positivity: (a) mean + 2.5 SD of young controls (age <28 years), (b) mean + 2.5 SD of elderly CU Aβ−, (c) Gaussian mixture modeling (GMM), and (d) the Youden index. Braak stages were assigned using seven (0 to VI) or four (0, I–II, III–IV, V–VI) categories. We evaluated inter‐ and intra‐tracer concordance (intra‐tracer, i.e., whether it follows the sequential Braak pattern). Results The intra‐tracer seven‐class Braak staging concordance ranged from 63% to 94%. With the highest intra‐tracer Braak concordance being achieved when using GMM cutoffs: 94% (MK6240) and 89% (Flortaucipir; Figure 1). Inter‐tracer agreement concordance ranged from 56% to 76%. The highest concordance emerged from the CU Elderly Aβ– cutoff optimizing the Braak II region for spill‐off (Figure 2). Using the Braak staging simplified version improved intra‐tracer concordance in both tracers (MK6240as well as inter‐tracer agreement (86.5%). Most inter‐tracer discrepancies were observed at Braak stages II–IV. Despite showing staging discordances, the distribution of cognitive status across the Braak stages is similar for both tracers (Figure 3). Conclusion These preliminary findings reveal some discrepancies in Braak staging when comparing MK6240 and Flortaucipir. Our results also suggest that adjustments in cutoffs and regions of interest can partially mitigate both inter‐ and intra‐tracer divergences. Finally, our analysis suggests robust concordance after adjustment and using 4 classes (0, I‐II, III‐IV, V‐VI).

Abstract Background Tau‐PET provides critical topographic information on tau deposition in the living brain. Integrating research‐based practices into clinical settings through qualitative methods, such as visual reads, is becoming increasingly important. This study aims to compare the performance of existing visual reading methods for [ 18 F]Flortaucipir (FTP) and [ 18 F]MK‐6240 (MK) and provide preliminary data toward a unified visual reading approach for all tau‐PET tracers using a head‐to‐head dataset. Methods The study design and development plans are illustrated in Figure 1. To evaluate previously published visual reading methods, two blinded raters conducted FTP and MK visual reads on 340 participants from the HEAD study. A unified visual reading method was developed and tested on 101 participants, using a composite of adjusted Braak regions to determine 3 stages of severity while harmonizing between tracers. For external validation, the method will be tested on four additional cohorts, each using one of the following tracers: FTP, MK, PI, or RO. Results Inter‐rater agreement using previous methods (Figure 2A) showed high Cohen’s kappa values (0.71‐0.77) for the low and high tau burden categories in both FTP and MK visual reads. However, agreement was substantially lower in the non‐AD‐like category, leading to overall agreement rates of 0.55 for FTP and 0.69 for MK. In contrast, the method developed here improved inter‐rater agreement across all categories (0.87‐0.96) and introduced a moderate tau (0.79‐0.83) burden group (Figure 2B), leading to higher overall agreement rates of 0.81 for FTP and 0.87 for MK (Figure 2C). For inter‐tracer agreement (Figure 3), previous methods resulted in discordant visual reads for FTP and MK in 34.4% of cases for rater 1 and 24.7% for rater 2. The developed method reduced this disagreement to 11.9% (rater 1) and 15.8% (rater 2), demonstrating improved consistency across different tracers. Conclusion This study demonstrates that previously published visual reading methods produced varying classifications depending on the tracer used. The method proposed here improved inter‐rater and inter‐tracer visual read agreements by developing a unified, tracer‐agnostic approach easy to be used. Importantly, this clinician‐friendly method has the potential for widespread adoption, offering a single harmonized visual reading technique for all tau‐PET tracers.

Abstract Background One of the hallmarks of Alzheimer’s disease (AD) is the progressive spread of tau pathology. However, heterogeneity in the spatial patterns of tau deposition at the individual level may contribute to distinct clinical presentations. Here, we investigated the different patterns of tau deposition across disease progression and the different cognitive profiles linked to these patterns. Method We investigated 106 tau PET positive individuals (16% cognitively unimpaired, 84% amnestic cognitively impaired) from the HEAD study using head‐to‐head Flortaucipir and MK6240 tau PET. For each tracer, we extracted the SUVR values from seven major brain networks: Limbic, Visual (Vis), Default Mode (Default), Dorsal Attention (DorsAttn), Frontoparietal, Somatomotor (SomMot) and Salience Ventral Attention (SalVentAttn) as well as the Entorhinal cortex. Pathways were identified by ranking regions from highest to lowest SUVR, with the region displaying the highest SUVR value being defined as the peak region. Individuals in late stages of tau deposition displaying distinct pathways were compared. Image averages and voxel‐wise group comparisons using tau negative individuals as the reference were generated for each pathway and tau PET tracer. Cognitive differences were assessed using ANOVA for MoCA score and domain‐specific composite scores for Memory, Attention and Visuospatial skills. Result In all individuals and tracers, tau in the Entorhinal cortex progressed to the Limbic network, supporting a common starting point for tau propagation. Beyond these initial stages, tau propagation followed three main distinct pathways following either the Vis, Default or DorsAttn networks (Figure 1). Voxel‐wise comparison between individuals across these three pathways captured clear and similar differences in tau PET uptake as measured with Flortaucipir and MK6240. (Figure 2). While MoCA and Memory scores did not differ between groups, individuals following the Vis network exhibited significantly lower Attention scores compared to those in the Default and DorsAttn networks. Additionally, individuals following the Vis network exhibited lower Visuospatial scores compared to the Default network group (Figure 3). Conclusion Our results indicate a model in which, although tau progresses hierarchically, its deposition peaks mainly follow specific networks, which drive tau deposition and, consequently, the clinical manifestations of AD.

Abstract Background HEAD is a longitudinal, multi‐site, non‐randomized study aiming to harmonize different tau PET tracers onto a common scale for enrichment and monitoring of anti‐tau clinical trials. To achieve this, the Uniτ scale is being developed using data from participants scanned head‐to‐head with Flortaucipir and MK6240, with a subset also scanned with PI2620 and RO948. However, access to the Uniτ scale has been limited to the study community and collaborators. To address this limitation, we have developed the Uniτ Ecosystem – a set of tools for tau PET harmonization designed to support research and clinical trials ( www.unitau.app ). Method The Uniτ Ecosystem was developed for multiple operating systems (Web/macOS/Windows/Ubuntu) (Figure 1) and currently include four tools. The Uniτ harmonization is an online ROI harmonization module that allows the user to harmonize SUVRs from spreadsheets. The Uniτ packages (R, Python and MATLAB) facilitates the transformation of SUVR values to the Uniτ scale using the most used programming languages in research settings. The Uniτ calculator app allows rapid and convenient harmonization. The Uniτ desktop app is a 3D harmonization module for voxel‐wise harmonization. The Uniτ parameters were calculated using a training set of 200 participants with head‐to‐head MK6240 and Flortaucipir, and 90 individuals with additional PI2620 and RO948. Researchers can use their own datasets and convert tau PET data to the Uniτ scale. Result The online ROI harmonization module allows users to harmonize tau PET SUVRs uploaded from spreadsheets and offers interactive visualizations such as scatter plots and histograms for tau positivity thresholds (Figure 2). The calculator app enables users to harmonize single SUVR values effortlessly. The desktop application for 3D tau PET harmonization leverages the functionalities of the previous modules to harmonize entire 3D images, eliminating the need for predefined ROIs. Additionally, this module provides visualization tools for the parametric Uniτ images in 2D and 3D (Figure 3). Uniτ tools never store any data on any external servers. Conclusion The Uniτ Ecosystem represents a significant advance in tau PET harmonization, providing a user‐friendly platform to harmonize and visualize tau PET data for research and clinical trials.

Abstract Background We showed that astrocyte reactivity, as measured by plasma GFAP levels, influences Aβ mediated tau PET pathology in cognitively unimpaired (CU) Aβ‐positive individuals. However, the link between GFAP and tau PET in individuals without detectable Aβ pathology remains elusive. The aim of the current study is to investigate the association between plasma GFAP and tau PET in CU Aβ PET‐negative individuals. Method We studied 147 CU Aβ PET‐negative participants from the HEAD cohort with plasma GFAP and p ‐tau217, as well as tau PET Flortaucipir and MK6240 data. Aβ positivity was determined by Aβ PET visual reading and Centiloid 12. Voxel‐wise linear regression models adjusted for age and sex tested the association of plasma GFAP and p ‐tau217 with tau PET. Further, the associations of plasma GFAP with peak tau PET SUVR values extracted from voxel‐wise association were fitted with a linear regression model adjusted for age and sex. Result Voxel‐wise analysis showed that plasma GFAP levels, but not plasma p ‐tau217 levels, were associated with tau PET in the medial temporal lobe (e.g., amygdala, entorhinal cortex, hippocampus) predominantly for Flortaucipir tau PET [Figure 1A, B, C, D]. The association between plasma p ‐tau217 and tau PET was weak in Aβ‐negative individuals. These results were similar when Aβ positivity was defined based on Centiloid 12. Furthermore, plasma GFAP and peak tau PET SUVR of Flortaucipir showed stronger association than that of MK6240 [Flortaucipir: β=0.4017, p <0.0001; MK6240: β=0.194, p = 0.0424; Figure 2A, B]. Conclusion We found an association between GFAP levels and tau PET uptake in individuals not expected to exhibit high levels of tau tangle‐related tracer uptake. Further analysis will be designed to elucidate the underpinning of this association, which could represent low levels of tau pathology, astrogliosis, or other factors.

Abstract Background Previous studies demonstrated an association between tau pathology and the development of neuropsychiatric symptoms (NPS) in individuals with Alzheimer’s disease (AD). However, the extent to which tau influences each specific NPS domain remains unclear. Here, we aim to investigate the association of tau and each NPS domain in the AD continuum. We hypothesized that tau plays a comparatively greater effect on the emergence of hyperactive and psychotic symptoms compared to other NPS domains. Method We assessed 385 individuals (216 cognitively unimpaired (CU), 128 MCI, and 41 AD dementia) from the HEAD study who underwent clinical assessments with the Neuropsychiatry Inventory Questionnaire (NPI‐Q) and had positron emission tomography (PET) for amyloid‐β (Aβ) ([ 18 F]AZD4694 or [ 11 C]PiB), and tau tangles ([ 18 F]MK6240) at the same visit. Tau SUVR values were tailored with a mask from Braak stages I‐VI, using the inferior cerebellar gray matter as a reference region. All individuals with dementia had a positive Aß‐PET. Voxel‐wise and Tobit censored regression tested the association between NPS domains and biomarkers accounting for age, sex, cognitive status, and study site. We used censored regression models to account for skewed data alongside a leave‐one‐out (loo) approach to identify which NPS domains most contributed to results. Result CI individuals had significantly higher NPI‐Q scores and Tau PET SUVR than CU (Figure 1A, 1B). NPI‐Q score was significantly associated with tau‐PET predominantly in fronto‐parietal regions. Removing irritability from the models strengthened this association (Figures 2A‐C). Regression loo models revealed that motor disturbances contributed most to the association between NPS and tau‐PET across all Braak stages. Notably, irritability had a negative effect on the association in all Braak stages, suggesting that tau does not play a role in the development of irritability and is highly associated with motor disturbances (Figure 3). Conclusion Our study suggests that tau contributes to the development of motor disturbances but has no effect on the development of irritability across the AD continuum. These findings provide additional rationale for the development of new therapeutics aiming to mitigate motor disturbances and irritability in AD patients.

Abstract Background Alzheimer's disease (AD) affects all brain cells and has complex genomic and immunological alterations. Previous research discovered missense AD risk or protective variants in microglial genes ABI3 and PLCG2 , respectively. Expression levels of these genes are altered in AD and can influence microglial function. This study aims to uncover protective, and risk microglial molecular signatures associated with these variants to determine their role in microglial subtypes and states in AD by single cell expression and functional studies. Method We generated microglia‐enriched snRNAseq data from donors harboring either AD protective PLCG2 or AD risk ABI3 missense mutations, or neither mutation. After standard snRNAseq QC, we performed differential expressed gene (DEG) analysis of all microglial cells between variant carriers and non‐carriers using MAST. We defined protective signature as genes that are both down in ABI3 and up in PLCG2 mutation‐carriers. In contrast, risk signature genes are up in ABI3 and down in PLCG2 mutation‐carriers. We investigated the conservation of protective and risk signatures across multiple datasets, including scRNAseq data from iPSC‐derived microglia cells carrying PLCG2 protective variant, snRNAseq data from AD‐resilient donors, and data from AD and other diagnostic groups across multiple brain regions sourced from external datasets. Result We obtained snRNAseq profiles of 35,000 microglia from AD variant carriers. Our DEG analysis among all microglia cells provided 227 microglial protective and 293 risk signature genes defined by these variants. Using integrated analysis of multiple internal and external datasets, we further narrowed down these signatures. We determined that these high‐confidence protective signature genes are downregulated in early AD, upregulated in late AD brains and positively‐correlated with protective variant load in in vitro models. In contrast, risk signature genes are upregulated in early AD, downregulated in late AD and resilient donors. Risk signature expression is decreased with protective PLCG2 variant load, and altered with Aβ treatment in in vitro models. Conclusion Our study uncovers microglia specific protective and risk signatures associated with AD using sn/scRNAseq datasets from multiple sources and models. These findings nominate novel immune targets and pathways with implications for microglial function in health and disease, and ultimately therapeutic potential.