Multi-omic characterization of brain changes in the vascular endothelial growth factor family during aging and Alzheimer's disease

An orchestra of machine learning methods reveals landmarks in single-cell data exemplified with aging fibroblasts

In this work, a Python framework for characteristic feature extraction is developed and applied to gene expression data of human fibroblasts. Unlabeled feature selection objectively determines groups and minimal gene sets separating groups. ML explainability methods transform the features correlating with phenotypic differences into causal reasoning, supported by further pipeline and visualization tools, allowing user knowledge to boost causal reasoning. The purpose of the framework is to identify characteristic features that are causally related to phenotypic differences of single cells. The pipeline consists of several data science methods enriched with purposeful visualization of the intermediate results in order to check them systematically and infuse the domain knowledge about the investigated process. A specific focus is to extract a small but meaningful set of genes to facilitate causal reasoning for the phenotypic differences. One application could be drug target identification. For this purpose, the framework follows different steps: feature reduction (PFA), low dimensional embedding (UMAP), clustering ((H)DBSCAN), feature correlation (chi-square, mutual information), ML validation and explainability (SHAP, tree explainer). The pipeline is validated by identifying and correctly separating signature genes associated with aging in fibroblasts from single-cell gene expression measurements: PLK3, polo-like protein kinase 3; CCDC88A, Coiled-Coil Domain Containing 88A; STAT3, signal transducer and activator of transcription-3; ZNF7, Zinc Finger Protein 7; SLC24A2, solute carrier family 24 member 2 and lncRNA RP11-372K14.2. The code for the preprocessing step can be found in the GitHub repository https://github.com/AC-PHD/NoLabelPFA, along with the characteristic feature extraction https://github.com/LauritzR/characteristic-feature-extraction.

Association of Neurotrophic Factors at Midlife With In Vivo Measures of β-Amyloid and Tau Burden 15 Years Later in Dementia-Free Adults

BACKGROUND AND OBJECTIVES

Neurotrophic factors (NTFs) play an important role in Alzheimer disease (AD) pathophysiology. Brain-derived neurotrophic factor (BDNF) and vascular endothelial growth factor (VEGF) are important NTFs. However, a direct link of BDNF and VEGF circulating levels with in vivo measures of amyloid-β (Aβ) and tau burden remains to be elucidated. We explored the relationship of BDNF and VEGF serum levels with future brain Aβ and tau pathology in a cohort of cognitively healthy, predominantly middle-aged adults and tested for possible effect modifications by sex and menopausal status.

METHODS

This cross-sectional analysis was conducted using data from the Framingham Heart Study (FHS), a community-based cohort study. The study sample included cognitively healthy participants from the FHS Offspring and Third-generation cohorts. BDNF and VEGF were measured in the third-generation cohort during examination cycles 2 (2005-2008) and 1 (2002-2005), respectively, and in the offspring cohort during examination cycle 7 (1998-2001). Participants underwent 11C-Pittsburgh compound B amyloid and 18F-Flortaucipir tau-PET imaging (2015-2021). Linear regression models were used to assess the relationship of serum BDNF and VEGF levels with regional tau and global Aβ, adjusting for potential confounders. Interactions with sex and menopausal status were additionally tested.

RESULTS

The sample included 414 individuals (mean age = 41 ± 9 years; 51% female). Continuous measures of BDNF and VEGF were associated with tau signal in the rhinal region after adjustment for potential confounders (β = -0.15 ± 0.06, p = 0.018 and β = -0.19 ± 0.09, p = 0.043, respectively). High BDNF (≥32,450 pg/mL) and VEGF (≥488 pg/mL) levels were significantly related to lower rhinal tau (β = -0.27 ± 0.11, p = 0.016 and β = -0.40 ± 0.14, p = 0.004, respectively) and inferior temporal tau (β = -0.24 ± 0.11, p = 0.028 and β = -0.26 ± 0.13, p = 0.049, respectively). The BDNF-rhinal tau association was observed only among male individuals. Overall, BDNF and VEGF were not associated with global amyloid; however, high VEGF levels were associated with lower amyloid burden in postmenopausal women (β = -1.96 ± 0.70, p = 0.013, per 1 pg/mL).

DISCUSSION

This study demonstrates a robust association between BDNF and VEGF serum levels with in vivo measures of tau almost 2 decades later. These findings add to mounting evidence from preclinical studies suggesting a role of NTFs as valuable blood biomarkers for AD risk prediction.

One-pot method for preparing DNA, RNA, and protein for multiomics analysis

Typical multiomics studies employ separate methods for DNA, RNA, and protein sample preparation, which is labor intensive, costly, and prone to sampling bias. We describe a method for preparing high-quality, sequencing-ready DNA and RNA, and either intact proteins or mass-spectrometry-ready peptides for whole proteome analysis from a single sample. This method utilizes a reversible protein tagging scheme to covalently link all proteins in a lysate to a bead-based matrix and nucleic acid precipitation and selective solubilization to yield separate pools of protein and nucleic acids. We demonstrate the utility of this method to compare the genomes, transcriptomes, and proteomes of four triple-negative breast cancer cell lines with different degrees of malignancy. These data show the involvement of both RNA and associated proteins, and protein-only dependent pathways that distinguish these cell lines. We also demonstrate the utility of this multiomics workflow for tissue analysis using mouse brain, liver, and lung tissue.

Alzheimer's disease, aging, and cannabidiol treatment: a promising path to promote brain health and delay aging

Alzheimer's disease (AD) is the most common neurodegenerative disease characterized by progressive memory loss, neurodegeneration, and cognitive decline. Aging is one of the risk factors for AD. Although the mechanisms underlying aging and the incidence rate of AD are unclear, aging and AD share some hallmarks, such as oxidative stress and chronic inflammation. Cannabidiol (CBD), the major non-psychoactive phytocannabinoid extracted from Cannabis sativa, has recently emerged as a potential candidate for delaying aging and a valuable therapeutic tool for the treatment of aging-related neurodegenerative diseases due to its antioxidant and anti-inflammation properties. This article reviews the relevant literature on AD, CBD treatment for AD, cellular senescence, aging, and CBD treatment for aging in recent years. By analyzing these published data, we attempt to explore the complex correlation between cellular senescence, aging, and Alzheimer's disease, clarify the positive feedback effect between the senescence of neurocytes and Alzheimer's disease, and summarize the role and possible molecular mechanisms of CBD in preventing aging and treating AD. These data may provide new ideas on how to effectively prevent and delay aging, and develop effective treatment strategies for age-related diseases such as Alzheimer's disease.

Hippocampus Glutathione S Reductase Potentially Confers Genetic Resilience to Cognitive Decline in the AD-BXD Mouse Population

Alzheimer's disease (AD) is a prevalent and costly age-related dementia. Heritable factors account for 58-79% of variation in late-onset AD, but substantial variation remains in age-of- onset, disease severity, and whether those with high-risk genotypes acquire AD. To emulate the diversity of human populations, we utilized the AD-BXD mouse panel. This genetically diverse resource combines AD genotypes with multiple BXD strains to discover new genetic drivers of AD resilience. Comparing AD-BXD carriers to noncarrier littermates, we computed a novel quantitative metric for resilience to cognitive decline in the AD-BXDs. Our quantitative AD resilience trait was heritable and genetic mapping identified a locus on chr8 associated with resilience to AD mutations that resulted in amyloid brain pathology. Using a hippocampus proteomics dataset, we nominated the mitochondrial glutathione S reductase protein (GR or GSHR) as a resilience factor, finding that the DBA/2J genotype was associated with substantially higher GR abundance. By mapping protein QTLs (pQTLs), we identified synaptic organization and mitochondrial proteins coregulated in trans with a cis-pQTL for GR. We found four coexpression modules correlated with the quantitative resilience score in aged 5XFAD mice using paracliques, which were related to cell structure, protein folding, and postsynaptic densities. Finally, we found significant positive associations between human GSR transcript abundance in the brain and better outcomes on AD-related cognitive and pathology traits in the Religious Orders Study/Memory and Aging project (ROSMAP). Taken together, these data support a framework for resilience in which neuronal antioxidant pathway activity provides for stability of synapses within the hippocampus.

A Continuous Extension of Gene Set Enrichment Analysis Using the Likelihood Ratio Test Statistics Identifies Vascular Endothelial Growth Factor as a Candidate Pathway for Alzheimer's Disease via ITGA5

BACKGROUND

Alzheimer's disease (AD) involves brain neuropathologies such as amyloid plaque and hyperphosphorylated tau tangles and is accompanied by cognitive decline. Identifying the biological mechanisms underlying disease onset and progression based on quantifiable phenotypes will help understand disease etiology and devise therapies.

OBJECTIVE

Our objective was to identify molecular pathways associated with hallmark AD biomarkers and cognitive status, accounting for variables such as age, sex, education, and APOE genotype.

METHODS

We introduce a pathway-based statistical approach, extending the gene set likelihood ratio test to continuous phenotypes. We first analyzed independently each of the three phenotypes (amyloid-β, tau, cognition) using continuous gene set likelihood ratio tests to account for covariates, including age, sex, education, and APOE genotype. The analysis involved 634 subjects with data available for all three phenotypes, allowing for the identification of common pathways.

RESULTS

We identified 14 pathways significantly associated with amyloid-β; 5 associated with tau; and 174 associated with cognition, which showed a larger number of pathways compared to biomarkers. A single pathway, vascular endothelial growth factor receptor binding (VEGF-RB), exhibited associations with all three phenotypes. Mediation analysis showed that among the VEGF-RB family genes, ITGA5 mediates the relationship between cognitive scores and pathological biomarkers.

CONCLUSIONS

We presented a new statistical approach linking continuous phenotypes, gene expression across pathways, and covariates like sex, age, and education. Our results reinforced VEGF RB2's role in AD cognition and demonstrated ITGA5's significant role in mediating the AD pathology-cognition connection.

A Continuous Extension of Gene Set Enrichment Analysis using the Likelihood Ratio Test Statistics Identifies VEGF as a Candidate Pathway for Alzheimer's disease

Background

Alzheimer's disease involves brain pathologies such as amyloid plaque depositions and hyperphosphorylated tau tangles and is accompanied by cognitive decline. Identifying the biological mechanisms underlying disease onset and progression based on quantifiable phenotypes will help understand the disease etiology and devise therapies.

Objective

Our objective was to identify molecular pathways associated with AD biomarkers (Amyloid-β and tau) and cognitive status (MMSE) accounting for variables such as age, sex, education, and APOE genotype.

Methods

We introduce a novel pathway-based statistical approach, extending the gene set likelihood ratio test to continuous phenotypes. We first analyzed independently each of the three phenotypes (Amyloid-β, tau, cognition), using continuous gene set likelihood ratio tests to account for covariates, including age, sex, education, and APOE genotype. The analysis involved a large sample size with data available for all three phenotypes, allowing for the identification of common pathways.

Results

We identified 14 pathways significantly associated with Amyloid-β, 5 associated with tau, and 174 associated with MMSE. Surprisingly, the MMSE outcome showed a larger number of significant pathways compared to biomarkers. A single pathway, vascular endothelial growth factor receptor binding (VEGF-RB), exhibited significant associations with all three phenotypes.

Conclusions

The study's findings highlight the importance of the VEGF signaling pathway in aging in AD. The complex interactions within the VEGF signaling family offer valuable insights for future therapeutic interventions.