Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2023 Jan 6:15:948456.
doi: 10.3389/fnmol.2022.948456. eCollection 2022.

Cell type-specific histone acetylation profiling of Alzheimer's disease subjects and integration with genetics

Easwaran Ramamurthy  1 , Gwyneth Welch  2 , Jemmie Cheng  2 , Yixin Yuan  1 , Laura Gunsalus  1 , David A Bennett  3 , Li-Huei Tsai  2   4 , Andreas R Pfenning  1
Affiliations

Cell type-specific histone acetylation profiling of Alzheimer's disease subjects and integration with genetics

Easwaran Ramamurthy et al. Front Mol Neurosci. .

Abstract

We profile genome-wide histone 3 lysine 27 acetylation (H3K27ac) of 3 major brain cell types from hippocampus and dorsolateral prefrontal cortex (dlPFC) of subjects with and without Alzheimer's Disease (AD). We confirm that single nucleotide polymorphisms (SNPs) associated with late onset AD (LOAD) show a strong tendency to reside in microglia-specific gene regulatory elements. Despite this significant colocalization, we find that microglia harbor more acetylation changes associated with age than with amyloid-β (Aβ) load. In contrast, we detect that an oligodendrocyte-enriched glial (OEG) population contains the majority of differentially acetylated peaks associated with Aβ load. These differential peaks reside near both early onset risk genes (APP, PSEN1, PSEN2) and late onset AD risk loci (including BIN1, PICALM, CLU, ADAM10, ADAMTS4, SORL1, FERMT2), Aβ processing genes (BACE1), as well as genes involved in myelinating and oligodendrocyte development processes. Interestingly, a number of LOAD risk loci associated with differentially acetylated risk genes contain H3K27ac peaks that are specifically enriched in OEG. These findings implicate oligodendrocyte gene regulation as a potential mechanism by which early onset and late onset risk genes mediate their effects, and highlight the deregulation of myelinating processes in AD. More broadly, our dataset serves as a resource for the study of functional effects of genetic variants and cell type specific gene regulation in AD.

Keywords: Alzheimer’s disease; Epigenomics; brain cell types; genetics; genomics.

PubMed Disclaimer

Conflict of interest statement

AP served as a paid consultant for Cognition Therapeutics, Inc., during the preparation of the manuscript. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Figures

FIGURE 1
FIGURE 1
FANS sorting captures neurons, microglia and oligodendrocyte enriched populations from postmortem brain tissue. (A) Workflow for sorting nuclei and performing H3K27ac ChIP-seq from postmortem human brain tissue: nuclei were isolated from fresh frozen hippocampus or prefrontal cortex and FANS was performed to collect NeuN+, Pu.1+, and NeuN-/Pu.1- populations. H3K27ac ChIP-seq was performed on each population (B). Genome browser visualization of H3K27ac signal over background (Input) averaged across all profiled samples for the three populations. Loci containing the genes RBFOX3 (NeuN), SPI1 (Pu.1) and OLIG2 (an oligodendrocyte marker) are visualized (C). Top heatmap displaying average H3K27ac enrichment at the promoters of marker genes (<5kb from TSS) from 15 cell type clusters profiled in Habib et al. (2017). Rows represent individual tissue samples. Columns represent the 15 different cell type clusters and are repeated three times to display NeuN+ specificity, Pu.1+ specificity and NeuN-/Pu.1- specificity. Bottom collapsed version of the top heatmap created by averaging the log2fc values for groups of samples defined by Aβ load, sex and brain region. Habib et al. (2017) cell type cluster abbreviations are defined here: exPFC, glutamatergic neurons from the PFC; GABA, GABAergic interneurons; exCA1/3, pyramidal neurons from the hippocampus CA region; exDG, granule neurons from the hippocampus dentate gyrus region; ASC, astrocytes; MICROGLIA, microglia; OLIGO, oligodendrocytes; OPC, oligodendrocyte precursor cells; NSC, neuronal stem cells; END, endothelial cells.
FIGURE 2
FIGURE 2
AD associated SNPs derived from GWAS prefer to colocalize with peaks enriched in the microglial population relative to peaks enriched in the OEG and neuronal populations (A,B). Results of stratified LD score regression from two AD GWAS studies (Jansen et al., 2019; Kunkle et al., 2019) and cell type-specific H3K27ac peaks. Plots show the estimated LD score regression coefficient for the three peak sets. Benjamini Hochberg FDR corrected q-values across the three tests for enrichment are indicated above each bar (C). Cell type enrichment of peaks annotated to sentinel SNPs at AD risk loci identified by Jansen et al. (2019) and Kunkle et al. (2019). Plots show fold change (log2-transformed) of H3K27ac signal for each population against the other two populations for (i) in black: peaks closest to the sentinel SNP at each locus associated with AD from GWAS, and (ii) in red: promoter peaks of early onset AD risk genes (APP, PSEN1, PSEN2). *Indicates DeSeq2 FDR q < 0.05 for the cell type-specific contrast. Sentinel SNPs that introduce missense mutations in proteins or SNPs where the closest H3K27ac peak is annotated > 1kb away are not included. This restriction was to ensure the analysis comprised only of SNPs that likely have functional effects on promoters or enhancer activity (D–F) top: Genome browser tracks of (i) reproducible peaks in each cell type for subjects without Aβ load, (ii) average H3K27ac signal in subjects without Aβ load for each cell type, and (iii) Manhattan plots of Jansen et al. (2019) and Kunkle et al. (2019) genetic variants. Plots are focused at loci where sentinel non-coding SNPs overlap with peaks enriched in non-neuronal cell types (d. INPP5D, e. BIN1, f. PICALM); bottom: zoomed in versions of the genome browser tracks displayed on top. INPP5D locus: the sentinel SNP rs10933431 overlaps with a peak that is enriched only in the microglial population; BIN1 locus: the top two AD-associated SNPs based on GWAS p-value (rs4663105 and rs6733839) overlap with peaks enriched in both the microglial and OEG populations; PICALM locus: the top two SNPs (rs10792832 and rs3851179) also overlap with non-neuronal peaks. Regions of overlap are highlighted with a yellow box.
FIGURE 3
FIGURE 3
OEG display the strongest acetylation differences associated with Aβ pathology, including peaks annotated to genes associated with EOAD and LOAD risk (A). Heatmap displaying number of significantly hyperacetylated (log2fc > 0, FDR q < 0.05) and significantly hypoacetylated peaks (log2fc < 0, FDR q < 0.05) from each brain region, sex, and cell type-specific contrast (B) left: Heatmap of normalized acetylation levels at 1962 H3K27ac peaks that were significantly hypoacetylated in AD female hippocampus OEG samples. Rows represent the 1,962 DARs and columns represent hippocampal OEG samples. Aβ load for each sample is indicated at the top of the heatmap. Right: A heatmap of the 1,962 peaks in male hippocampal OEG samples is included for comparison. DARs annotated to EOAD and LOAD risk genes are labeled in red and black, respectively. Peaks near STMN4 and MYRF are annotated in green (C). Heatmap of the 1,029 H3K27ac peaks that were significantly hyperacetylated in AD dlPFC OEG samples. Peaks annotated to EOAD and LOAD risk genes are labeled in red and black, respectively. The ADAMTS18 promoter-proximal peak is annotated in green (D). Distance to TSS distribution of (i) 1,962 OEG female hippocampus hypoacetylated DARs, (ii) 1,029 OEG dlPFC hyperacetylated DARs and (iii) the full consensus set of peaks (E). Enrichment heatmap of top gene ontology terms for 6 peak sets (1) 1,962 OEG female hippocampus hypoacetylated DARs (2) 1,029 OEG dlPFC hyperacetylated DARs (3) all other Aβ associated DARs (4) neuron, (5) microglia, and (6) OEG cell type-specific hyperacetylated peaks. Color intensity represents hypergeometric fold enrichment in number of peaks over background (full consensus peak set), *indicates FDR q < 0.05, **indicates FDR q < 0.01.
FIGURE 4
FIGURE 4
EOAD and LOAD risk genes exhibit epigenomic and transcriptomic perturbations in oligodendrocytes (A–I). Genome browser tracks displaying average H3K27ac signal in OEG samples from subjects with and without Aβ load (yellow and blue tracks, respectively). Regions displayed include EOAD and LOAD risk loci, as well as differentially acetylated regions near ADAMTS18 and MYRF (J) RT-qPCR of select genes annotated to DARs identified in AD OEG female hippocampus. Panel shows violin plots of gene expression measured by RT-qPCR in hippocampal Olig2+ nuclei collected from an independent cohort of AD and non-AD subjects. q-values for differential expression between high Aβ and low+mid Aβ subjects are indicated on top for each gene. Correction was applied across the 9 tests (K) left panel: comparison with existing snRNA-seq from AD dlPFC (Mathys et al., 2019) reveals an average increase in gene expression near hyperacetylated regions in OEG dlPFC. Violin plots depict log2fc values from differential expression analysis between AD and non-AD subjects in oligodendrocytes (Mathys et al., 2019). These log2fc values are derived from 500 genes annotated to the OEG dlPFC hyperacetylated DARs that reside in putative promoters (<5kb from TSS). Log2fc violin plots are shown for two different contrasts performed in Mathys, Valderrain et al.: no pathology vs. pathology and no pathology vs. early pathology. Q-values from t-tests (null hypothesis: mean log2fc = 0, alternate hypothesis: mean log2fc > 0) are reported for the two violin plots. Correction was applied across the two tests. Right panel: Specific genes associated with OEG dlPFC hyperacetylated DARs display increased transcription in AD. Individual log2fc values are shown. TSS distance cutoffs were not used for this right panel. FDR q-values from the differential expression analysis for each gene are also provided for both contrasts.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. Alsema A. M., Jiang Q., Kracht L., Gerrits E., Dubbelaar M. L., Miedema A., et al. (2020). Profiling microglia from AD donors and non-demented elderly in acute human post-mortem cortical tissue. bioRxiv [Preprint]. 10.1101/2020.03.18.995332 - DOI - PMC - PubMed
    1. Bartzokis G. (2011). Alzheimer’s disease as homeostatic responses to age-related myelin breakdown. Neurobiol. Aging 32 1341–1371. 10.1016/j.neurobiolaging.2009.08.007 - DOI - PMC - PubMed
    1. Bekris L. M., Yu C. E., Bird T. D., Tsuang D. W. (2010). Genetics of Alzheimer Disease. J. Geriatr. Psychiatry Neurol. 23:213. 10.1177/0891988710383571 - DOI - PMC - PubMed
    1. Bennett D. A., Buchman A. S., Boyle P. A., Barnes L. L., Wilson R. S., Schneider J. A. (2018). Religious Orders Study and Rush Memory and Aging Project. J. Alzheimers Dis. 64 S161–S189. 10.3233/JAD-179939 - DOI - PMC - PubMed
    1. Bennett D. A., Schneider J. A., Buchman A. S., Barnes L. L., Boyle P. A., Wilson R. S. (2012a). Overview and findings from the rush Memory and Aging Project. Curr. Alzheimer Res. 9 646–663. 10.2174/156720512801322663 - DOI - PMC - PubMed
  NODES
Association 1
INTERN 1
Project 2
twitter 2