The Key Role of DNA Methylation in Alzheimer's Disease | Alzheimer's & Dementia

Alzheimer's disease (AD) is a progressive neurodegenerative disease characterized by progressive memory impairment, cognitive impairment, personality changes, and language disorders. The pathological and physiological mechanisms of AD are complex, and the pathogenesis is still unclear. Its pathological characteristics are β- Starch like (Aβ) plaques and neurofibrillary tangles (NFT). In recent years, researchers have been exploring the molecular mechanisms of AD neuropathology, developing various prevention and detection methods β significant progress has been made in areas such as plaques and NFT methods, but the etiology of typical Alzheimer's disease, especially late onset AD (LOAD), remains unclear to this day.

The fragility of brain regions is a significant feature of the neuropathological development of AD, among which the parahippocampal gyrus (PHG) region is one of the earliest affected and most vulnerable brain regions. Recent studies have shown that PHG (as well as the entorhinal cortex (ER)) is the region with the greatest changes in the epigenome and transcriptome of Alzheimer's disease. Therefore, studying the multi omics of PHG may reveal the key molecular mechanisms of Alzheimer's disease.

Recently, a research team from Mount Sinai Icahn School of Medicine in the United States published an article titled "Genome wide metabolic regulation of multiscale gene networks in Alzheimer's disease" in Alzheimer's & Dementa. The research team identified 270 differentially methylated regions (DMRs) associated with AD clinical, pathological features, and genome-wide gene expression changes, and further defined a gene level total methylation score (OMS) to quantify the impact of DMR on each gene and protein.?This study provides a framework for future multidimensional data integration and may discover new targets for AD drug development.

The research team first analyzed genome-wide methylation changes related to AD neuropathology, and analyzed the PHG region tissues of 201 postmortem control groups, mild cognitive impairment (MCI), and AD brain samples from the Mount Sinai Brain Bank (MSBB) cohort. After strict quality control, 776,508 CpG sites were retained for downstream analysis out of 196 PHG samples.

The research team analyzed the association between methylomics differences and mean plaque density (PLQ-Mn), which is one of the key criteria for quantifying AD neuropathology. 13,755 differential methylation sites (DMP) were identified between AD and healthy normal control (NL) brains, with approximately 70% of DMP located near genomic features, indicating their potential regulatory role in transcription.?It is worth noting that genes near highly methylated DMP are enriched in neuronal and synaptic functions, which are often downregulated in AD; the genes near low methylation DMP are enriched in immune system processes and other pathways, which are usually activated in AD. The above results indicate that changes in methylation levels in AD patients are related to function.

To understand the spatial organization of methylated CpG sites in AD, the research team used the sliding window algorithm (comb-p) to identify adjacent CpG site regions with high correlation, and defined DMR as a genomic region. The results showed that a total of 270 obvious DMR were found, with chromosomes 1 and 11 having the most (27 DMR) and chromosomes 9 and 21 having the least (2 DMR); there is a significant correlation between the average methylation level of DMR and PLQ Mn.

DNA methylation is one of the key components of post transcriptional regulation of gene expression. The research team investigated the impact of DMR as a whole on the mRNA or protein expression changes of each gene, and quantified the contribution of all DMR to each gene using variance decomposition analysis (Figure 2). The results showed that an average of 34% of the expression level changes of 23,201 mRNA were related to DMR. Interestingly, DMR has a greater impact on protein expression variation, explaining an average of 39% of protein expression variation.

To summarize the genes affected by methylation, the research team ranked the genes based on the methylation contribution obtained from variance decomposition analysis. A total of 3,449 potential methylation regulatory genes (MRG) and 2,632 potential methylation regulatory proteins (MRP) were identified.?Among 1,345 common genes, 601 underwent significant changes in AD and NL, with the majority (589) expressing changes in the same direction as NL in AD and participating in functional pathways and biological processes such as neural and synaptic responses, as well as immune responses.

Due to the complex interactions between DMR and gene expression characteristics, as well as between different DMR, the research team has developed a new statistical method called "OMS" to quantify the net methylation effect on gene and protein expression. Negative OMS (OMS [-]) represents the net effect of decreased gene expression caused by methylation, while positive OMS (OMS [+]) represents the net effect of increased gene expression caused by methylation; OMS (0) indicates no methylation effect; the OMS of a module is the sum of the mRNA-OMS of all genes in that module.

The research team calculated the OMS at the gene and gene module levels, and evaluated the correlation between gene module OMS, module correlation score (RS), and AD. mRNA-OMS and protein-OMS were used to represent the overall impact of methylation on gene expression at the mRNA and protein levels. The results showed a significant correlation between module RS and module mRNA-OMS (-), as well as between module RS and module mRNA-OMS (+), indicating that methylation plays a crucial role in regulating the co expression of AD neuropathological basic genes.

In addition, the research team analyzed the epigenetic regulation of key driver genes (KD) in the gene network and ranked the key driver genes (KDG) in the previously predicted Bayesian causal gene network in the PHG region. The results showed that out of 1454 KDG, 774 were associated with DMR, of which 517 were mRNA OMS (-) and 257 were mRNA OMS (+). When divided into positive and negative groups, KDG scores were significantly correlated with their mRNA-OMS. Furthermore, the research team detected the methylation levels of downstream genes of KDG and found that KDG scores were associated with mRNA-OMS of downstream genes. The above results indicate the profound role of methylation in regulating the primary network KD and its downstream genes.

Finally, the research team evaluated the impact of methylation on gene expression using two methods (variance decomposition and direct correlation) and observed significant overlap in the gene expression features obtained by the two methods (Figure 3). For example, in the MSBB queue, approximately 50% of mRNA features and over 60% of protein features obtained through variance decomposition analysis appeared in the gene list of direct correlation analysis, and similar results were also observed in the independent validation queue ROSMAP. These methods work together to demonstrate that DNA methylation has a regulatory effect on the expression of a large number of genes in the human brain.

In summary, the research team first demonstrated the presence of changes in DNA methylation levels related to AD, further identifying DMR, revealing methylomics changes related to AD neuropathology, and developing a new indicator OMS to quantify the impact of methylation on each gene and protein. This study provides a new method for analyzing the relationship between DNA methylation and gene/protein expression, and emphasizes the importance of epigenetic mechanisms in human diseases such as AD.

?

Reference:

Erming Wang, Minghui Wang, Lei Guo,?et al.?Genome-wide methylomic regulation of multiscale gene networks in Alzheimer's disease.?Alzheimer's?&?Dementia?(2023).