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
Meta-Analysis
. 2020 Jul 14;32(2):107908.
doi: 10.1016/j.celrep.2020.107908.

Meta-Analysis of the Alzheimer's Disease Human Brain Transcriptome and Functional Dissection in Mouse Models

Ying-Wooi Wan  1 Rami Al-Ouran  2 Carl G Mangleburg  1 Thanneer M Perumal  3 Tom V Lee  4 Katherine Allison  4 Vivek Swarup  5 Cory C Funk  6 Chris Gaiteri  7 Mariet Allen  8 Minghui Wang  9 Sarah M Neuner  10 Catherine C Kaczorowski  10 Vivek M Philip  10 Gareth R Howell  10 Heidi Martini-Stoica  11 Hui Zheng  12 Hongkang Mei  13 Xiaoyan Zhong  13 Jungwoo Wren Kim  14 Valina L Dawson  15 Ted M Dawson  16 Ping-Chieh Pao  17 Li-Huei Tsai  17 Jean-Vianney Haure-Mirande  18 Michelle E Ehrlich  19 Paramita Chakrabarty  20 Yona Levites  20 Xue Wang  21 Eric B Dammer  22 Gyan Srivastava  23 Sumit Mukherjee  3 Solveig K Sieberts  3 Larsson Omberg  3 Kristen D Dang  3 James A Eddy  3 Phil Snyder  3 Yooree Chae  3 Sandeep Amberkar  24 Wenbin Wei  25 Winston Hide  26 Christoph Preuss  10 Ayla Ergun  27 Phillip J Ebert  28 David C Airey  28 Sara Mostafavi  29 Lei Yu  7 Hans-Ulrich Klein  30 Accelerating Medicines Partnership-Alzheimer’s Disease Consortium  31 Gregory W Carter  10 David A Collier  32 Todd E Golde  20 Allan I Levey  33 David A Bennett  7 Karol Estrada  27 T Matthew Townsend  34 Bin Zhang  9 Eric Schadt  9 Philip L De Jager  30 Nathan D Price  6 Nilüfer Ertekin-Taner  35 Zhandong Liu  36 Joshua M Shulman  37 Lara M Mangravite  38 Benjamin A Logsdon  39
Affiliations
Meta-Analysis

Meta-Analysis of the Alzheimer's Disease Human Brain Transcriptome and Functional Dissection in Mouse Models

Ying-Wooi Wan et al. Cell Rep. .

Abstract

We present a consensus atlas of the human brain transcriptome in Alzheimer's disease (AD), based on meta-analysis of differential gene expression in 2,114 postmortem samples. We discover 30 brain coexpression modules from seven regions as the major source of AD transcriptional perturbations. We next examine overlap with 251 brain differentially expressed gene sets from mouse models of AD and other neurodegenerative disorders. Human-mouse overlaps highlight responses to amyloid versus tau pathology and reveal age- and sex-dependent expression signatures for disease progression. Human coexpression modules enriched for neuronal and/or microglial genes broadly overlap with mouse models of AD, Huntington's disease, amyotrophic lateral sclerosis, and aging. Other human coexpression modules, including those implicated in proteostasis, are not activated in AD models but rather following other, unexpected genetic manipulations. Our results comprise a cross-species resource, highlighting transcriptional networks altered by human brain pathophysiology and identifying correspondences with mouse models for AD preclinical studies.

Keywords: Alzheimer's disease; RNA-seq; aging; coexpression analysis; differential expression analysis; meta-analysis; mouse models; neuroinflammation; transcriptome.

PubMed Disclaimer

Conflict of interest statement

Declaration of Interests The authors declare no competing interests.

Figures

Figure 1.
Figure 1.. Human Consensus RNA-Seq Coexpression Modules
(A) Gene set overlap (Fisher’s exact test p value) was examined among 30 AD-associated coexpression modules, highlighting five consensus clusters (A, B, C, D, and E). (B) Coexpression modules were evaluated using expression-weighted cell type enrichment analysis based on human single-cell RNA-seq. (C) Coexpression modules were examined for overlap (Fisher’s exact test) with curated AD gene sets from GeneCards, Panther, the Database of Genotypes and Phenotypes (dbGaP), IGAP, Online Mendelian Inheritance in Man (OMIM), Biocarta, Wikipathways, and the Kyoto Encyclopedia of Genes and Genomes (KEGG). We also evaluated overlap with coexpression modules derived from the constituent cohorts, including oligodendroglial modules identified by Mayo and MSSM, module 109 from ROSMAP, and an RNA-binding protein rich module from Emory. (D) Coexpression module enrichment for AD susceptibility gene candidates from GWAS, based on MAGMA. Consensus cluster B modules appear strongly enriched for AD risk loci.
Figure 2.
Figure 2.. Cross-Species Study Design and Data
(A) Analytic design for examining overlaps between 30 Alzheimer’s disease (AD)-associated human coexpression modules and differentially expressed gene sets from 376 experimental comparisons in mouse models. (B) Ninety-six mouse studies were selected based on relevance to AD and other neurodegenerative disorders. Distribution of keywords is shown among all studies. RNA-seq was reprocessed using a standard pipeline. See Table S3 for details on all included mouse studies. (C) The differentially expressed gene sets represent mouse models of AD, Huntington’s disease (HD), frontotemporal dementia-amyotrophic lateral sclerosis (FTD-ALS), spinocerebellar ataxia 1 (SCA1), Rett syndrome (RETT), Parkinson’s disease (PD), Creutzfeldt-Jakob disease (CJD), neurofibromatosis (NF), or other neurodegenerative mechanisms. A t-distributed stochastic neighbor embedding (t-SNE) plot including all mouse differential expression signatures highlight both disease-specific and overlapping features among heterogeneous neurodegenerative models. See also Figure S3.
Figure 3.
Figure 3.. Overview of Human-Mouse Overlaps and Concordance
Heatmaps show overlap (top) and concordance (bottom) among 30 human coexpression modules (rows) and 251 sets of differentially expressed genes (DEGs; columns) from mouse model comparisons. The average sample size is 8.4 (range 4–28 total samples). Mouse-human overlap significance, calculated using the hypergeometric test, is represented in grayscale (−log10[padj]). Direction (red/blue) and extent of concordance (intensity) for gene expression changes are also indicated (bottom). The color bar at the top annotates all DEGs based on whether they derive from Alzheimer’s disease (AD) models (pink), other neurodegenerative models (purple), or other experimental manipulations potentially relevant to AD mechanisms (orange). The color barat left denotes cluster membership (A–E). Cluster E modules (brown, asterisk) show sparse overlap with AD mouse model DEGs. See Tables S4, S5, and S6 for details on all experimental comparisons and comprehensive results.
Figure 4.
Figure 4.. Human Coexpression Module Overlaps with AD Mouse Models
(A) Heatmaps show overlap (top, hypergeometric test) and concordance (bottom) among human coexpression modules and sets of differentially expressed genes (DEGs) from Alzheimer’s disease (AD) mouse models. Mouse-human overlap significance, calculated using the hypergeometric test, is represented in grayscale (−log10[padj]). The color bar at the top denotes APP (pink), MAPT (orange), or other (purple) model comparisons. The estimated pathologic burden (plaques/tangles and neuronal loss) in APP and MAPT models is also annotated in green. Cluster E modules (brown, asterisk) shows parse overlap with AD mouse models. Selected overlaps are denoted as follows: squares, APP models with oligodendrocyte-enriched module overlaps; cross-hatches, CRND8-APP models showing sustained activation of microglial modules from 6 months onward; arrowhead, transient activation of neuronal modules in TG4510-MAPT model preceding microglial module overlap; circles, co-activation of neuronal and microglial modules. See Tables S4 and S5 for details on all experimental comparisons, including sample sizes and genotypes, along with comprehensive results. See Figure S4 for duplicated panel including detailed model annotations. See Figure S6 for additional analysis of overlap specificity. (B) Representative overlaps of human modules with mouse DEGs. The hypergeometric test was applied to assess significance. Gene counts are noted in black, including for overlapping and non-overlapping regions. To assess concordance between human brains and mouse models, gene counts are shown, noting increased (red) or decreased expression (blue), including for the human coexpression module and the overlapping mouse gene set. M147 was derived from Srinivasan et al. (2016). (C) Mouse model overlaps highlight age- and sex-dependent changes. Increasing (red) or decreasing (blue) gene expression and magnitude of changes shown as overlap (%) between the mouse DEG set and module. Cell type module clusters are denoted by colors at panel bottom, as in (A). (D) Enrichment of human sex-specific DEGs from random-effects meta-analysis among coexpression modules. Consensus cluster C shows downregulation in AD among females.
Figure 5.
Figure 5.. Overlaps with Other Mouse Models
(A) Heatmaps show overlap among human coexpression modules and sets of differentially expressed genes (DEGs) from mouse models, including pure aging, neurodegenerative disorders, and other experimental manipulations. HD, Huntington’s disease; FTD-ALS, frontotemporal dementia-amyotrophic lateral sclerosis; SCA1, spinocerebellar ataxia 1; CJD, Creutzfeldt-Jakob disease. Mouse-human overlap significance, calculated using the hypergeometric test, is represented in grayscale (−log10[padj]). Overlaps between HD model expression signatures (Langfelder et al., 2016) and neuronal gene-enriched human coexpression modules recapitulate polyglutamine length (M100, Q92 versus M94, Q175) and brain region dependence (M100/M94, striatum versus M72/M81, cortex). Other manipulations generate signatures similar to AD models, including PTCH1 knockout (M183) (Ung et al., 2018), nmf205 (M182) (Ishimura et al., 2016), and neuroserpin mutant (M205) (Guadagno et al., 2017). Modules poorly enriched for cell type signatures (asterisk, right) show selected overlaps with FTD-ALS models (M28, M43; Ibrahim et al., 2013, and Lagier-Tourenne et al., 2013, respectively) and other, unexpected genetic manipulations (M158, M111, M54, and M156; Vied et al., 2016, Maze et al., 2015, Narayanan et al., 2014, and Holmes et al., 2016, respectively). See Tables S5 and S6 for comprehensive results, including sample sizes for all comparisons. See Figure S5 for comprehensive heatmaps representing overlaps with HD, FTD-ALS, SCA1, and aging models. (B) Representative overlaps of human modules with mouse DEGs, as in Figure 4B. The hypergeometric test was applied to assess significance.

References

    1. Adalbert R, Nogradi A, Babetto E, Janeckova L, Walker SA, Kerschensteiner M, Misgeld T, and Coleman MP (2009). Severely dystrophic axons at amyloid plaques remain continuous and connected to viable cell bodies. Brain 132, 402–416. - PubMed
    1. Adamcsek B, Palla G, Farkas IJ, Derenyi I, and Vicsek T (2006). CFinder: locating cliques and overlapping modules in biological networks. Bioinformatics 22, 1021–1023. - PubMed
    1. Ahn Y-Y, Bagrow JP, and Lehmann S (2010). Link communities reveal multiscale complexity in networks. Nature 466, 761–764. - PubMed
    1. Allen M, Carrasquillo MM, Funk C, Heavner BD, Zou F, Younkin CS, Burgess JD, Chai H-S, Crook J, Eddy JA, et al. (2016). Human whole genome genotype and transcriptome data for Alzheimer’s and other neurodegenerative diseases. Sci. Data 3, 160089. - PMC - PubMed
    1. Allen M, Wang X, Burgess JD, Watzlawik J, Serie DJ, Younkin CS, Nguyen T, Malphrus KG, Lincoln S, Carrasquillo MM, et al. (2018). Conserved brain myelination networks are altered in Alzheimer’s and other neurodegenerative diseases. Alzheimers Dement. 14, 352–366. - PMC - PubMed

Publication types

Grants and funding