Michal Schnaider Beeri

Spontaneous language offers a scalable, noninvasive approach for identifying biological indicators of preclinical Alzheimer's disease (AD). We examined prediction of APOE4 carrier status and cerebral amyloid-beta burden using Cookie Theft picture descriptions from two cohorts of Hebrew- and English-speaking participants (IRAP, N=134; WRAP, N=1,541). Amyloid-beta burden was classified using lowest-quartile, median, and highest-quartile cutoffs, capturing accumulation below conventional amyloid-positivity thresholds. We compared embeddings from pretrained large language models (LLMs) with standard cognitive test scores, reflecting conventional clinical assessment, and theory-driven linguistic features, reflecting traditional language analysis. Embedding-based models achieved the highest F1 scores among constituent models in five of eight tasks involving APOE4 status and global amyloid-beta burden. High prediction disagreement suggested complementary signals across approaches. An any-positive ensemble significantly outperformed every constituent model and a prior-informed random baseline across all eight tasks, and significantly outperformed all constituent models in 21 of 24 regional amyloid-beta tasks. These findings highlight the complementarity of LLM-based, theory-driven, and cognitive measures for predicting AD-related genetic risk and amyloid pathology, and support further investigation of their potential for early screening approaches.

Mapping genetic effects on distal genes and regulatory programs is essential for understanding how genetic variation influences complex diseases; yet trans -QTL discovery remains limited by weak effect sizes, extensive multiple testing and poor reproducibility. Here, we introduce TransMap , a generalizable computational framework that combines multi-scale trans -QTL mapping with multi-phenotype genetic colocalization to improve the discovery and interpretation of distal regulatory variants. Beyond individual trans -eGenes, TransMap constructs local co-regulatory program phenotypes that account for background expression variation to improve the detection of weak, coordinated distal genetic effects. Across a broad range of simulation settings, TransMap achieved 9.2-36.7% higher trans -QTL discovery power than existing methods while maintaining well-controlled type I error rates. We applied TransMap to bulk and single-nucleus transcriptomic profiles from the aging human brain in the ADSP Functional Genomics Consortium (FunGen-AD), further integrating proteomic, glycoproteomic and metabolomic biomarker measurements (average N= 476 donors). Genome-wide fine-mapping of TransMap single-gene and program-level QTL yielded variant annotations with 2.75-2.88x heritability enrichment across 57 complex diseases and traits (average N= 307K), with complementary heritability information captured across regulatory scales. Multi-phenotype colocalization identified 49,532 trans -regulatory events supported by shared genetic signals with cis -eQTL or biomarker-QTL, including 1,825 candidate cis–trans–biomarker regulatory configurations. TransMap further revealed an asymmetric cross-cell-type regulatory architecture in which neuronal cis -regulatory signals preferentially associate with coordinated glial trans -regulatory responses, supported by co-culture experiments and spatial transcriptomics. Alzheimer’s disease (AD) GWAS-prioritized colocalization identified trans -regulatory associations in 115 AD-associated loci, including a KNOP1 cis -eQTL associated in trans with a microglial regulatory program containing MS4A4E , with additional support from statistical mediation analysis. TransMap provides a multi-scale, cell-type-resolved atlas of trans -regulatory variation and a generalizable framework for elucidating the regulatory architecture of complex diseases.

INTRODUCTION: Type 2 diabetes (T2D) is associated with cognitive decline, but the role of APOE ε4 – a known Alzheimer's risk allele – in cognition among admixed populations with T2D remains unclear. METHODS: We analyzed 883 Brazilian adults with T2D (median age 68 years) from primary care, excluding those with dementia. Cognitive function was assessed using Mini‐Mental State Examination (MMSE), and APOE genotypes were determined. MMSE errors were modeled using negative binomial regression. RESULTS: APOE alleles showed no association with MMSE errors overall. Older age and self‐reported Black/Brown race predicted more errors, whereas higher education and physical activity predicted fewer. Sensitivity analyses indicated a possible inverse association between APOE ε4 and MMSE errors among self‐reported Black/Brown participants. DISCUSSION: APOE ε4 was not associated with cognition in this cohort. The suggestive protective effect observed in Black/Brown participants should be considered hypothesis‐generating, underscoring the need for further research in admixed populations with T2D. Highlights: APOE ε4 was not linked to cognitive errors in the full T2D Brazilian cohort. Trend toward fewer MMSE errors among self‐reported Black/Brown APOE ε4 carriers. Higher education, physical activity, and BMI were protective against MMSE errors. Black/Brown race and older age were associated with more MMSE errors. Results support multifactorial assessment and more studies in admixed T2D populations.

Type 2 diabetes (T2D) is consistently related to an increased risk of cognitive decline and dementia. However, the molecular underpinnings of this association remain poorly understood. In this study, we applied a novel mass spectrometry-based glycoproteomic methodology to profile serum glycoproteins in older adults with T2D, aiming to identify glycopeptiforms associated with cognitive impairment. Our method allowed comprehensive profiling of N glycosylation in addition to the unique ability to profile glycation events on specific amino acid sites. Serum samples from initially cognitively normal older adults with T2D were collected, with participants classified as cognitive decliners (who developed impairment) and non-decliners (who maintained normal cognition over time). We identified significant differences in the abundance of glycopeptiforms between these groups, noting that certain glycopeptiforms exhibited unique changes over time in decliners. We identified 13 glycopeptiforms that exhibited significant differences between the groups both at baseline and in their rates of change over time. Pathway analysis indicated that glycation events were linked to metabolic pathways while glycosylation to immune-related pathways, aligning with established links between these processes and cognitive decline. This study offers new insights into glycoproteoform alterations in older adults with T2D experiencing cognitive decline. It highlights the potential of specific glycopeptiforms as biomarkers for early cognitive impairment in T2D. Further validation in larger cohorts will enhance our understanding of glycosylation and glycation in T2D and potentially lead to the discovery of novel treatment targets for T2D-related cognitive decline. Raw data and search are available via ProteomeXchange with identifier PXD050780.

Objective: High BMI, which poorly represents specific fat depots, is linked to poorer cognition and higher dementia risk, with different associations between sexes. This study examined associations of abdominal fat depots with cognition and brain volumes and whether sex modifies this association.Methods: A total of 204 healthy middle‐aged offspring of Alzheimer's dementia patients (mean age = 59.44, 60% females) underwent abdominal magnetic resonance imaging to quantify hepatic, pancreatic, visceral, and subcutaneous adipose tissue and to assess cognition and brain volumes.Results: In the whole sample, higher hepatic fat percentage was associated with lower total gray matter volume (β = −0.17, p < 0.01). Primarily in males, higher pancreatic fat percentage was associated with lower global cognition (males: β = −0.27, p = 0.03; females: β = 0.01, p = 0.93) executive function (males: β = −0.27, p = 0.03; females: β = 0.02, p = 0.87), episodic memory (males: β = −0.28, p = 0.03; females: β = 0.07, p = 0.48), and inferior frontal gyrus volume (males: β = −0.28, p = 0.02; females: β = 0.10, p = 0.33). Visceral and subcutaneous adipose tissue was inversely associated with middle frontal and superior frontal gyrus volumes in males and females.Conclusions: In middle‐aged males at high Alzheimer's dementia risk, but not in females, higher pancreatic fat was associated with lower cognition and brain volumes. These findings suggest a potential sex‐specific link between distinct abdominal fat with brain health.

Background: The blood-brain barrier (BBB) is composed of brain microvasculature that provides selective transport of solutes from the systemic circulation into the central nervous system to protect the brain and spinal microenvironment. Damage to the BBB in the acute phase following traumatic brain injury (TBI) is recognized as a major underlying mechanism leading to secondary long-term damage. However, due to the lack of technological ability to detect subtle BBB disruption (BBBd) in the chronic phase, the presence of chronic BBBd is disputable. Thus, the dynamics and course of long-term BBBd post-TBI remains elusive. Methods: 30 C57BL/6 male mice subjected to TBI using our weight drop closed head injury model and 19 naïve controls were scanned by MRI up to 540 days following injury. BBB maps were calculated from delayed contrast extravasation MRI (DCM) with high spatial resolution and high sensitivity to subtle BBBd, enabling depiction and quantification of BBB permeability. At each time point 2-6 animals were sacrificed and their brains were extracted, sectioned, and stained for BBB biomarkers including: blood microvessel coverage by astrocyte using GFAP, AQP4, ZO-1 gaps and IgG leakage. Results: we found that DCM provided depiction of subtle yet significant BBBd up to 1.5 years following TBI, with significantly higher sensitivity than standard contrast-enhanced T1-weighted and T2-weighted MRI (BBBd volumes main effect DCM/T1/T2 p<0.0001 F(2,70)=107.3, time point p<0.0001 F(2,133, 18.66)=23.53). In 33% of the cases, both in the acute and chronic stages, there was no detectable enhancement on standard T1-MRI, nor detectable hyper-intensities on T2-MRI, whereas DCM showed significant BBBd volumes. BBBd values of TBI mice at the chronic stage were fonud significantly higher compered to age matched naïve animals at 30,60 and 540 days. The calculated BBB maps were histologically validated by determining significant correlation between the calculated levels of disruption and a diverse set of histopathological parameters obtained from different brain regions, presenting different components of the BBB. Summary and Discussion: Cumulative evidence from recent years points to BBBd as a central component of the pathophysiology of TBI. Therefore, it is expected that routine use of highly sensitive non-invasive techniques to measure BBBd, such as DCM with advanced analysis methods, may enhance our understanding of the changes in BBB function following TBI. Application of the DCM technology to other CNS disorders, as well as to normal aging, may shed light on the involvement of chronic subtle BBBd in these conditions.

Objective: Self‐rated health (SRH) is a predictor for poor health outcomes and cognition. Older adults with type 2 diabetes mellitus (T2D) have multi‐morbidity and greater cognitive impairment. In the present study we investigated the association of SRH with cognitive decline and brain pathology in older adults with T2D.Methods: Participants (n = 1122) were from the Israel Diabetes and Cognitive Decline study, and SRH was categorised as low (n = 202), moderate (n = 400) or high (n = 520). Cognition was measured by four cognitive domains: episodic memory, executive functions, language, and attention/working memory. Global cognition was the average of the cognitive domains. Statistical models adjusted for sociodemographic, cardiovascular, and clinical variables. In a randomly selected subsample (n = 230) that had magnetic resonance imaging, we examined relationships between baseline SRH and brain characteristics (white matter hyperintensities [WMHs], hippocampal, and total grey matter [GM] volumes).Results: Low SRH was associated with a decline in executive functions, which accelerated over time when compared to high SRH (est = −0.0036; p = p = 0.009). Low SRH at baseline was associated with higher volumes of WMHs (est = 9.8420; p < 0.0008). SRH was not associated with other cognitive domains, or with hippocampal and total GM.Conclusions: Low SRH is associated with cognitive decline in T2D older adults and may serve as a risk assessment. WMHs may represent an underlying mechanism.

Background: Post‐mortem human brain tissue is the gold‐standard for neurodegenerative disease diagnosis and is invaluable for elucidating cellular and molecular mechanisms. Therefore, neurodegenerative disease brain banks are a critical component of research to advance diagnosis and develop effective disease modifying therapies. The Israel National Brain Repository (INBR) was established in 2018 at the Sheba Medical Center via a collaboration with the Mount Sinai Neuropathology Brain Bank in New York. Israel is a unique country with a diverse population including substantial populations of individuals of European descent, Middle‐Eastern North African (MENA) descent, and combinations of both.Method: INBR predominantly recruits from two prospective longitudinal cohort studies whose participants are at high dementia risk: older adults with type 2 diabetes (the Israel Diabetes and Cognitive Decline study) and middle‐aged offspring of people with Alzheimer’s Disease (The Israel Registry for Alzheimer’s Prevention). Participants in both cohorts are phenotyped for cognition, lifestyle, brain imaging, blood biomarkers, and genetics. Biospecimen collection and processing protocols are aligned with other Neurobiobanks in the USA (see flowchart). Additionally, all participants in both cohorts are members of the second largest healthcare provider in Israel, the Maccabi Healthcare Services, permitting linkage to longitudinal medical records dating from 1998.Result: N = 450 participants have signed intent to autopsy. Among the N = 285 with race/ethnicity data available, 30% are MENA. 14 brain and 6 spinal cord donations have been completed. At time of death, donors had an average age of 81.1 years, 10 had normal cognition, 3 had dementia, and one had amyotrophic lateral sclerosis.Conclusion: The potential strengths of the INBR is in the population characteristics of Israel, the ethnic diversity and admixture, the linkage to 25‐years of medical records data, and the small size of the country allowing donations to be received from anywhere in the country. The main challenge we face is cultural barriers to brain donation that exist among some segments of Israeli society. The ultimate goal is to capitalize on the strengths and uniqueness of INBR to develop novel therapies for dementia.