Much like Laurel and Hardy, Bert and Ernie, and Pinky and the Brain, neurons and glia are firm friends and are seldom encountered in isolation. The more numerous glia support and protect their neural chums, helping them to conduct nerve impulses; however, their close companionship and disparate numbers can lead to analytical problems!
Research into epigenetic dysregulation during aging can promote the development of Alzheimer’s disease (AD) typifies this problem; two recent epigenome-wide association studies (EWAS) on human brain tissues highlighted tissue composition heterogeneity as a significant source of “noise” in epigenetic profiles (Study 1 and 2). This noise weakens the distinction of age, cell-type, and disease-related changes and diminishes the strength of any conclusions.
To remove noise from the EWAS equation, researchers (and best buddies, we hope) from the lab of Jörn Walter (University of Saarland, Saarbrücken, Germany) broke-up the brain cell friendship by sorting neuronal and non-neuronal nuclei from postmortem human brain tissues before analysis employing the Illumina HumanMethylation 450k bead array platform. NeuN immunolabeling sorted the cells into a positive fraction (neurons) and a negative fraction (mostly glia) from 31 fresh frozen brain tissue samples, thereby generating 62 epigenome-wide cell-type-specific profiles.
Here is what the amiable team of scientists discovered in this first of a kind study in a small cohort:
- In the context of aging, cell-sorting permitted the identification of genes with cell-type-specific methylation signatures and highlighted differential methylation dynamics
- Aging generally leads to DNA hypomethylation in both neurons and glia
- Aging-related differentially methylated CpGs (DMCGs) associate with genes such as CLU (a well-known AD risk gene), FAM53B, and NCOR2 in neurons and RAI1, ELOVL2, and FHL2 in glia
- Interestingly, the CLU-associated hypomethylation event may lead to the expression of a pro-apoptotic isoform in older subjects
- There exists only a small overlap of the top 1000 aging-related DMCGs between neurons and glia, indicating distinct methylation dynamics upon aging; however, some CpGs display ‘synchronized’ dynamics across neurons and glia, including the pan-aging marker ELOVL2
- In the context of disease, the authors uncovered neuron- and glia-specific associations of DNA methylation with the different stages of AD pathology (Braak stages) that are mostly different from those discovered for aging
- Similar to aging, analysis indicates that neuronal and glial cells respond differentially during AD progression
- DMCGs exist at genes involved in neurotransmitter homeostasis and transport in neurons, and genes such as DIP2C, GNG2, and ANK1 in glia
- DMCGs in the regulatory regions of the anion channel LRRC8B gene occur in both neurons and glia, suggesting a link to AD development
- Finally, a meta-analysis of a combination of the cell-type-specific results (which should reflect the combined epigenetic burden at the different stages of AD) emphasized methylation changes associated with the HOXA3 and ANK1 genes
- Comparisons to other EWAS analyses uncovered the age-independent hypomethylation of the critical AD gene APP during AD progression in neurons and glia
Overall, this study emphasizes that while a duo such as neurons and glia may work well together, their epigenomes respond differently in the context of aging and disease. Furthermore, the authors also advise cell sorting before EWAS analysis as a means to generate robust conclusions.
For more on how separating cellular best buddies can aid our understanding of epigenetic dysregulation in Alzheimer’s Disease, see Epigenetics & Chromatin, July 2018.