A core transcriptional signature of human microglia: derivation and 1
utility in describing region-dependent alterations associated with 2
Alzheimer’s disease 3
Anirudh Patir
1
, Barbara Shih
1
, Barry W. McColl
1,2
and Tom C. Freeman
1†
4
5
1. The Roslin Institute, University of Edinburgh, Easter Bush, Midlothian, 6
Scotland, UK EH25 9RG. 7
2. UK Dementia Research Institute at The University of Edinburgh, Edinburgh 8
Medical School, The Chancellor's Building, 49 Little France Crescent, 9
Edinburgh, EH16 4TJ, UKUK. 10
11
†
Corresponding author 12
13
Running title: A functional profile of human microglia. 14
15
Acknowledgments 16
T.C.F. and B.W.M. are funded by an Institute Strategic Programme Grant funding 17
from the Biotechnology and Biological Sciences Research Council [BB/J004227/1]. 18
B.W.M. receives funding from the UK Dementia Research Institute and Medical 19
Research Council [MR/L003384/1]. B.S. is supported by Experimental Medicine 20
Challenge Grant funding from the Medical Research Council [MR/M003833/1]. 21
22
Conflict of Interest Statement 23
The authors have no competing financial interests. 24
25
Word Count: 5781 26
27
.CC-BY-NC-ND 4.0 International licensea
certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under
The copyright holder for this preprint (which was notthis version posted April 26, 2018. ; https://doi.org/10.1101/308908doi: bioRxiv preprint
Abstract 28
Growing recognition of the pivotal role microglia play in neurodegenerative and 29
neuroinflammatory disorders has accentuated the need to better characterize their 30
function in health and disease. Studies in mouse, have applied transcriptome-wide 31
profiling of microglia to reveal key features of microglial ontogeny, functional profile 32
and phenotypic diversity. Whilst similar in many ways, human microglia exhibit clear 33
differences to their mouse counterparts, underlining the need to develop a better 34
understanding of the human microglial profile. On examining published microglia 35
gene signatures, little consistency was observed between studies. Hence, we set out 36
to define a conserved microglia signature of the human central nervous system 37
(CNS), through a comprehensive meta-analysis of existing transcriptomic resources. 38
Nine datasets derived from cells and tissue, isolated from different regions of the 39
CNS across numerous donors, were subjected independently to an unbiased 40
correlation network analysis. From each dataset, a list of coexpressing genes 41
corresponding to microglia was identified. Comparison of individual microglia clusters 42
showed 249 genes highly conserved between them. This core gene signature 43
included all known markers and improves upon published microglial signatures. The 44
utility of this signature was demonstrated by its use in detecting qualitative and 45
quantitative region-specific alterations in aging and Alzheimer’s disease. These 46
analyses highlighted the reactive response of microglia in vulnerable brain regions 47
such as the entorhinal cortex and hippocampus, additionally implicating pathways 48
associated with disease progression. We believe this resource and the analyses 49
described here, will support further investigations in the contribution of human 50
microglia towards the CNS in health and disease. 51
52
53
54
55
56
57
58
.CC-BY-NC-ND 4.0 International licensea
certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under
The copyright holder for this preprint (which was notthis version posted April 26, 2018. ; https://doi.org/10.1101/308908doi: bioRxiv preprint
Keywords 59
Microglia, transcriptome, neurodegenerative disease, aging, Alzheimer’s. 60
Table of Contents: Main points 61
• Published microglial transcriptional signatures in mouse and human show 62
poor consensus. 63
• A core transcriptional signature of human microglia with 249 genes was 64
derived and found conserved across brain regions, encompassing the CNS. 65
• The signature revealed region-dependent microglial alterations in Alzheimer’s, 66
highlighting susceptible CNS regions and the involvement of TYROBP 67
signaling. 68
69
.CC-BY-NC-ND 4.0 International licensea
certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under
The copyright holder for this preprint (which was notthis version posted April 26, 2018. ; https://doi.org/10.1101/308908doi: bioRxiv preprint
Introduction 70
Microglia are the most abundant myeloid cell type in the central nervous system 71
(CNS), accounting for approximately 5-20% of the brain parenchyma depending on 72
region (Lawson, Perry, Dri, & Gordon, 1990; Mittelbronn, Dietz, Schluesener, & 73
Meyermann, 2001). These cells are phenotypically plastic and exhibit a wide 74
spectrum of activity influenced by local and systemic factors (Cunningham, 2013; 75
Perry & Holmes, 2014). Through development into adulthood, microglia influence the 76
proliferation and differentiation of surrounding cells while regulating processes such 77
as myelination, synaptic organization and synaptic signaling (Colonna & Butovsky, 78
2017; Hoshiko, Arnoux, Avignone, Yamamoto, & Audinat, 2012; Paolicelli et al., 79
2011; Prinz & Priller, 2014). As the primary immune sentinels of the CNS, microglia 80
migrate towards lesions and sites of infection, where they attain an activated state 81
that reflects their inflammatory environment (Leong & Ling, 1992). In these states, 82
they can support tissue remodeling and phagocytose cellular debris, toxic protein 83
aggregates and microbes (Colonna & Butovsky, 2017; Li & Barres, 2017). During 84
neuroinflammation these cells coordinate an immune response by releasing 85
cytokines, chemoattractants and presenting antigens, thereby communicating with 86
other immune cells locally and recruited from the circulation (Hanisch & Kettenmann, 87
2007; Hickey & Kimura, 1988; Scholz & Woolf, 2007). 88
In common with mononuclear phagocyte populations throughout the body, recent 89
studies have begun to reveal the diversity of microglial phenotypes in health, aging 90
and disease states, as well as their unique molecular identity in relation to other CNS 91
resident cells and non-parenchymal macrophages (Durafourt et al., 2012; Hanisch, 92
2013; Li & Barres, 2017; McCarthy; Salter & Stevens, 2017). The application of 93
transcriptomic methods has been integral to these advances by enabling an 94
unbiased and panoramic perspective of the functional profile of microglia. In addition 95
to an improved understanding of the variety of context-dependent microglial 96
phenotypes, other key benefits have arisen from these studies, notably the 97
development of new tools to label, isolate and manipulate microglia (Bennett et al., 98
2016; Butovsky et al., 2014; Hickman et al., 2013; Satoh et al., 2016). Although most 99
studies have been conducted in mice, a considerable body of data is now emerging 100
from human post-mortem and biopsy tissue (Darmanis et al., 2015; Galatro et al., 101
2017; Gosselin et al., 2017; Olah et al., 2018; Y. Zhang et al., 2016). Whilst there are 102
.CC-BY-NC-ND 4.0 International licensea
certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under
The copyright holder for this preprint (which was notthis version posted April 26, 2018. ; https://doi.org/10.1101/308908doi: bioRxiv preprint
many conserved features between rodent and human microglia, the importance of 103
further refining our understanding specifically of human microglia is underscored by 104
important differences that have been observed between them (Butovsky et al., 2014; 105
Galatro et al., 2017; Miller, Horvath, & Geschwind, 2010). 106
Recent transcriptomic studies have sought to characterize the human microglial 107
transcriptomic signature from the CNS of non-neuropathologic individuals using data 108
derived from either cells or tissue isolated from different brain regions (Darmanis et 109
al., 2015; Galatro et al., 2017; Hawrylycz et al., 2012; Oldham et al., 2008). These 110
analyses have been crucial in expanding our knowledge of their functional biology, 111
however, our preliminary analyses found there to be little inter-study agreement 112
across the published microglia gene signatures. Such inconsistency may have arisen 113
due to technical differences in tissue sampling, brain areas analyzed, differences in 114
patient characteristics and biological variance including the heterogeneity of different 115
microglia populations (Grabert et al., 2016; Lai, Dhami, Dibal, & Todd, 2011; Lawson 116
et al., 1990; Vincenti et al., 2016; Yokokura et al., 2011). This highlighted a need to 117
derive a refined human microglial signature that would enable a more precise 118
characterization of these cells in the healthy and diseased human brain. We 119
therefore set out to define the core transcriptional signature of human microglia, i.e. 120
shared by all microglial populations of the human CNS. To achieve this, we have 121
performed an extensive meta-analysis of nine human cell and tissue transcriptomics 122
datasets derived from numerous brain regions and donors. Secondly, we have used 123
this signature to investigate region-dependent changes, while highlighting the 124
influence of microglial numbers and activation in human tissue transcriptomics for 125
Alzheimer’s and aging. 126
Methods 127
Comparison of published microglial signatures 128
Ten publications that defined microglial signatures, four in human and six in mouse, 129
were identified (Table 1). To compare across studies, genes from each signature 130
were converted to a common identifier i.e. HGNC (Povey et al., 2001) or MGI (Shaw, 131
2009) for human and mouse, respectively, using the online tool g:Profiler (Reimand 132
et al., 2016). Subsequently, the tool was also used for interspecies comparison 133
based on the MGI homology database, identifying human orthologues of mouse 134
.CC-BY-NC-ND 4.0 International licensea
certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under
The copyright holder for this preprint (which was notthis version posted April 26, 2018. ; https://doi.org/10.1101/308908doi: bioRxiv preprint
