Microglia Heterogeneity in the Single-Cell Era

Cell Reports

Review Microglia Heterogeneity in the Single-Cell Era Takahiro Masuda,1 Roman Sankowski,1,2 Ori Staszewski,1,2 and Marco Prinz1,3,4,* 1Institute

of Neuropathology, Faculty of Medicine, University of Freiburg, Freiburg, Germany for Clinician Scientists, Faculty of Medicine, University of Freiburg, Freiburg, Germany 3Signalling Research Centres BIOSS and CIBSS, University of Freiburg, Freiburg, Germany 4Center for Basics in NeuroModulation (NeuroModulBasics), Faculty of Medicine, University of Freiburg, Freiburg, Germany *Correspondence: [email protected] https://doi.org/10.1016/j.celrep.2020.01.010 2Berta-Ottenstein-Programme

Microglia are resident immune cells in the central nervous system (CNS) that are capable of carrying out prominent and various functions during development and adulthood under both homeostatic and disease conditions. Although microglia are traditionally thought to be heterogeneous populations, which potentially allows them to achieve a wide range of responses to environmental changes for the maintenance of CNS homeostasis, a lack of unbiased and high-throughput methods to assess microglia heterogeneity has prevented the study of spatially and temporally distributed microglia subsets. The recent emergence of novel single-cell techniques, such as cytometry by time-of-flight mass spectrometry (CyTOF) and single-cell RNA sequencing, enabled scientists to overcome such limitations and reveal the surprising context-dependent heterogeneity of microglia. In this review, we summarize the current knowledge about the spatial, temporal, and functional diversity of microglia during development, homeostasis, and disease in mice and humans. Microglia are resident macrophages of the central nervous system (CNS), where they play crucial roles for the maintenance of homeostasis in cooperation with other CNS cells (Colonna and Butovsky, 2017; Prinz et al., 2019; Ransohoff and Perry, 2009). Microglia represent around 10% of adult CNS cells, with variable cell density across distinct CNS regions (Lawson et al., 1990). Recent fate-mapping studies clearly demonstrated that microglia originate from early erythromyeloid progenitors (EMPs) in the extraembryonic yolk sac, which seed the CNS of the embryo (Ginhoux et al., 2010; Gomez Perdiguero et al., 2015; Kierdorf et al., 2013; Schulz et al., 2012), differentiating into matured microglia through a stepwise developmental program (MatcovitchNatan et al., 2016). Remarkably, microglia self-renew through low-rate proliferation in combination with apoptosis throughout life (Ask3ew et al., 2017; Tay et al., 2017), without any contribution from bone-marrow-derived monocytes to the microglial pool, at least under homeostatic conditions (Bruttger et al., 2015). Historically, microglia have been distinguished based on their morphology, density, electrophysiological properties, and surface expression of a small panel of immune molecules (De Biase et al., 2017; Lawson et al., 1990; Schmid et al., 2002) and thus have been assumed to be a heterogeneous cell population within the CNS (Hanisch, 2013). Indeed, given the fact that microglia serve a variety of biologically important roles—such as supporting neuronal development, synaptic pruning, or scavenging for dying cells (Butovsky and Weiner, 2018)—it was thought for many years that specialized subsets of microglia already preexist in situ to achieve such a wide range of responses and functions in health. In addition, under disease conditions, they readily adapt their gene expressional profile and finally shift to reactive states on demand by virtue of their plasticity (Masuda and Prinz, 2016). However, until recently, the heterogeneity of microglia across different CNS regions during development, homeostasis, and perturbations had not been investigated on a single-cell

level. Furthermore, very little was known about context-dependent microglia subclusters during disease. The main reasons for those finite insights were major technical limitations that did not allow the proper spatial and temporal representation of microglia multiplicity at the single-cell level. Earlier single-cell analyses of microglia, such as cell flow cytometry, in situ hybridization, or immunohistochemistry, were limited to probing a few selected proteins or RNAs, which clearly hindered the ability to study comprehensive landscapes of microglial diversity. The recent advent of novel technologies, such as single-cell RNA sequencing (scRNA-seq), has enabled the profiling of single cells with high-throughput datasets, which offers an unbiased alternative workflow allowing the sequencing of cells without prior knowledge of genes of interest as well as the clustering of cells based on their transcriptional signatures. In addition, single-cell mass spectrometry (cytometry by time-of-flight [CyTOF]) with more than 40 different surface markers now also allows the study of complex cellular diversity on a larger scale. Such singlecell analysis can help build a more comprehensive molecular view of microglia, which makes it possible to identify novel markers, pathways, and regulatory factors that are critical for their function during development, health, and disease. In this review, we provide an overview of recent reports that explore microglial heterogeneity using novel advanced singlecell technologies (Table 1). These studies clearly facilitated our understanding of the microglia diversity during development, homeostasis, and perturbation in both mice and humans. Parenchymal Microglia and Non-Parenchymal CNSAssociated Macrophages as Distinct CNS Myeloid Cell Types There are two major, anatomically distinct resident macrophage populations in the CNS: microglia and CNS-associated macrophages (CAMs; also known as border-associated macrophages [BAMs]; Prinz et al., 2019). The latter are localized at the non-

Cell Reports 30, February 4, 2020 ª 2020 The Author(s). 1271 This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

Table 1. Overview of Recent Studies with scRNA-Seq or CyTOF Reference

Disease Model if Applicable (CNS Regions Tested)

Protocol Used

Age (CNS Regions Tested)

Matcovitch-Natan et al., 2016

MARS-seq

embryonic (E)10.5, E12.5, E14.5, E16.5, postnatal days (P)3, P6, P9, 8-week (whole brain), E10.5 and E12.5 (yolk sac)

N/A

Keren-Shaul et al., 2017

MARS-seq

1-, 3-, 6-, 8-month (whole brain), 6-month (cortex, cerebellum), 7-week and 20-month (whole brain)

5xFAD mice (cortex, cerebellum), mSOD1(G93A) ALS-model mice (spinal cord)

Mathys et al., 2017

SMART-seq2

adult (hippocampus)

CK-p25 mice (hippocampus)

Sousa et al., 2018

Drop-seq with minor changes

adult (whole brain)

LPS injection (whole brain)

Hammond et al., 2019

10x Chromium

E14.5, P4/5, P30, P100, P540 (whole brain)

lysolecithin-mediated demyelination (corpus callosum)

Li et al., 2019

SMART-seq2

E14.5 (whole brain), P7 and P60 (cortex, hippocampus, cerebellum, striatum, olfactory bulb, chroid plexus)

N/A

~o et al., 2019 Jorda

CEL-seq2

adult (spinal cord)

experimental autoimmune encephalomyelitis (spinal cord)

Masuda et al., 2019

SMART-seq2

E16.5 (forebrain, midbrain, cerebellum, spinal cord), 3W and 16-week (cortex, hippocampus, cerebellum, spinal cord, facial nucleus, corpus callosum)

cuprizone-mediated de-/remyelination (corpus callosum), facial nerve axotomy (facial nucleus)

Darmanis et al., 2015

Fluidigm C1 System

gestational 16-18-week (cortex), adult (cortex)

N/A

Zhong et al., 2018

SMART-seq2 with minor modification

gestational 8-26-week (prefrontal cortex)

N/A

Masuda et al., 2019

CEL-seq2

22-54-year (cortex: temporal lobe and frontal lobe)

multiple sclerosis (cortex: parietal lobe)

Sankowski et al., 2019

CEL-seq2

14-74-year (cortex: temporal lobe and frontal lobe)

IDH-wild-type glioblastoma

Mrdjen et al., 2018

8-week, 1.5-year (whole CNS)

APP/PS1 mice, experimental autoimmune encephalomyelitis (whole CNS),

Ajami et al., 2018

adult (whole CNS)

experimental autoimmune encephalomyelitis (whole CNS), R6/2 mice (whole CNS), mSOD1(G93A) ALS-model mice (whole CNS)

Bo¨ttcher et al., 2019

subventricular zone, thalamus, cerebellum, temporal robe, frontal lobe

N/A

Sankowski et al., 2019

14-74-year (cortex: temporal lobe and frontal lobe)

IDH-wild-type glioblastoma

scRNA-Seq: Mouse Microglia

scRNA-Seq: Human Microglia

CyTOF: Mouse Microglia

CyTOF: Human Microglia

parenchymal CNS interfaces such as the perivascular space, the meninges, and the choroid plexus (CP) (Kierdorf et al., 2019; Prinz et al., 2017; Prinz and Priller, 2014; Van Hove et al., 2019). Two remarkable properties that are shared by microglia and CAMs are their prenatal origin and longevity. Similar to microglia, CAMs—with the exception of CP macrophages that show a mixture of ontogeny (Van Hove et al., 2019)—arise from extraembryonic yolk sac EMPs, and they persist over a very long period of time with a low rate of self-renewal (Gold-

1272 Cell Reports 30, February 4, 2020

mann et al., 2016). In addition, microglia and CAMs share some marker genes such as the ionized calcium-binding adaptor molecule 1 (Iba-1; also known as allograft inflammatory factor 1 [Aif1]), the colony-stimulating factor 1 receptor (Csf1r), or the fractalkine receptor (Cx3cr1) (Ajami et al., 2018; Goldmann ~o et al., 2019; Mrdjen et al., 2018). Because et al., 2016; Jorda of these broad similarities, microglia and CAMs have historically been combined in conventional bulk analyses. However, recent studies with scRNA-seq analysis clearly segregated microglia

Figure 1. Dynamic Change of Microglial Heterogeneity during Development, Aging, and Perturbations Highly transcriptionally diverse microglia clusters exist at embryonic time points in mice, the extent of which gradually decreases over the course of development, resulting in limited heterogeneity of microglia during adulthood. In contrast, the diversity of microglia increases again during aging, whereas less is known about region-specific diversity in the aged brain. On the other hand, higher regional heterogeneity is observed at early postnatal time points, especially in white-matter regions such as the cerebellum or the corpus callosum. During disease conditions, microglia are able to once again change their transcriptional phenotypes and give rise to more heterogeneous substates in a context-dependent manner. Differentially colored cells indicate diverse microglia transcriptional states.

and CAMs by their distinct transcriptomic signature with the specification of molecular markers for each cell type. Tmem119 (transmembrane protein 119), P2ry12 (P2Y purinergic receptor 12), and Hexb (beta-hexosaminidase subunit beta) were defined as microglia-enriched genes, whereas CAMs were characterized by their high expression of genes including Ms4a7 (membrane-spanning 4-domains subfamily A member 7), Mrc1 (mannose receptor 1, C-type mannose receptor 1), and Pf4 ~o et al., 2019; Zei(platelet factor 4) (Goldmann et al., 2016; Jorda sel et al., 2015), highlighting distinct CNS-resident myeloid cell populations with discrete transcriptional profiles and potentially diverse functions during physiology and pathophysiology. Moreover, a recent study has reported the existence of a non-parenchymal myeloid cell population located on the CP epithelium that exhibits a typical microglia signature similar to ‘‘disease-associated microglia’’ (DAM) described before in a mouse model of Alzheimer’s disease (Keren-Shaul et al., 2017). This macrophage population shares the ontogeny with microglia (Van Hove et al., 2019), suggesting a unique homeostatic type of microglia localized in the CP. The specific function of this particular microglia population in the CP remains unknown. Microglia Diversity during Development and Homeostasis in Mice The CNS tissue constitutes a well-organized cellular network with multiple cell types such as neurons, astrocytes, oligodendrocytes, mural cells, CAMs, and microglia. Each cell type can be subdivided based on their transcriptional profiles (Zeisel et al., 2015). Such complexity of the cellular microenvironment of the CNS and the varied proportion of different cell types across the CNS regions may shape the characteristics of microglia in situ. Indeed, microglia are continuously sensing molecules of the surrounding microenvironment through their surface proteins (e.g., receptors or transporters), which could facilitate their diverse behaviors and functions (Abiega et al., 2016; Dissing-

Olesen et al., 2014; Fontainhas et al., 2011; Rothhammer et al., 2018). In fact, a large body of evidence has revealed the phenotypical diversity of microglia at bulk transcriptomic or proteomic levels in the CNS (Grabert et al., 2016; Sharma et al., 2015). Cerebellar microglia were shown to highly express F4/80, compared to microglia from other brain regions (Lawson et al., 1990). Further, expression levels of CD68 (a lysosome-associated membrane protein) in microglia were reported to differ across the CNS regions in rodents (Ayata et al., 2018; Doorn et al., 2015) as well as in humans (Mildner et al., 2017), suggesting regionally variable phagocytic capacity. In addition, microglial expression levels of Tnf (encoding tumor necrosis factor a) and Fcgr2 (encoding Fc gamma receptor type 2) varied between the CNS regions with high expression in the hippocampus (Ren et al., 1999). Finally, white-matter microglia were found to exhibit a higher expression of activating molecules such as Usp18 (ubiquitin-specific protease 18) when compared to cortical microglia (Goldmann et al., 2015). Although these findings have given important insights into microglial multiplicity and have thus facilitated the description of several microglia states, those expressional analyses with bulk RNA analyses hindered the ability to study microglia heterogeneity comprehensively in an unbiased fashion. Recent studies with single-cell analysis have therefore shed completely new light on the spatial and developmental heterogeneity of microglia and provided for the first time a high-resolution view of the transcriptional landscape of microglia subtypes across multiple regions of the murine CNS during development (Hammond et al., 2019; Keren-Shaul et al., 2017; Li et al., 2019; Masuda et al., 2019; Matcovitch-Natan et al., 2016; Figure 1). Of note, highly diverse microglia were observed at embryonic and early postnatal time points in mice where several novel subclasses of microglia were identified (Hammond et al., 2019; Li et al., 2019; Masuda et al., 2019; Matcovitch-Natan et al., 2016). Hammond et al. (2019) analyzed the transcriptomes of more than

Cell Reports 30, February 4, 2020 1273

76,000 individual microglia using the 103 technology and identified a unique microglia population that highly expresses Ms4a family members, such as Ms4a7—a recently identified marker ~o et al., 2019)—in the embryonic brain. This subfor CAMs (Jorda set seemed transcriptionally similar to that of CAMs but was otherwise clearly distinct (Hammond et al., 2019). Furthermore, the authors confirmed the existence of Ms4a7+ microglia in the brain parenchyma of embryos by in situ hybridization. Whether this small subset gives rise to CAMs or further differentiates into mature microglia remains to be elucidated. However, its functional relevance and that of other identified states in the developing embryonic brain need to be examined in the future. Notably, other recent studies did not identify the Ms4a7+ microglia subset in the brains of embryos (Li et al., 2019; Masuda et al., 2019), which might be due to the different sequencing technologies and clustering methods used. Additional embryonic subpopulations of microglia characterized by high expression of genes such as Apoe (encoding apolipoprotein E), Ctsb (encoding cathepsin b), or Tmsb4x (encoding thymosin beta-4) were identified (Masuda et al., 2019). Apoe is known to regulate the activity of microglia (Krasemann et al., 2017). The Ctsb+ microglia subpopulation also expressed a set of genes related to lysosomal functionality such as Ctsd (encoding cathepsin D) and Lamp1 (lysosomal-associated membrane protein 1) at a high level, suggesting enhanced lysosomal activity in this subclass, which might reflect their elevated phagocytic capacity. The role of Tmsb4x in microglia has not yet been determined. Notably, the functional relevance of all newly described microglia subpopulations during development remains unknown and needs to be investigated in the future. In a recent study, we showed variable proportions of respective clusters of microglia across CNS regions with no spatially restricted subpopulations at the embryonic stage (Masuda et al., 2019). The differential enrichment of microglial clusters across the regions during development may reflect distinct maturation stages of these regions. A distinct amoeboid microglia subset characterized by high expression of Clec7a (encoding C-type lectin domain family 7 member A), Gpmbp (encoding transmembrane glycoprotein NMB), Spp1 (encoding secreted phosphoprotein 1, also known as osteopontin), Lgals3 (encoding galectin-3), and Igf1 (encoding insulin-like growth factor 1) was frequently found in white-matter regions in early postnatal brain regions, such as the corpus callosum and cerebellum, but was almost absent during adulthood (Hagemeyer et al., 2017; Hammond et al., 2019; Li et al., 2019). Interestingly, many of the genes upregulated in this white-matter population found during early postnatal development are known to characterize activated adult microglia surrounding amyloid plaques in a mouse model of Alzheimer’s disease in an APOEand TREM2 (triggering receptor expressed on myeloid cells-2)dependent manner. This similarity with plaque-associated microglia called DAM (Keren-Shaul et al., 2017; Krasemann et al., 2017) led the authors to assume a similar molecular machinery underlying the phenotype of the postnatal microglia subset. However, the density of this subset in mice lacking either the Trem2 or Apoe gene was comparable to that of wild-type counterparts (Li et al., 2019), indicating TREM2- and APOE-independent mechanisms for the induction of this microglia type. The morphological characteristics and the crucial functionality of

1274 Cell Reports 30, February 4, 2020

this microglia subset in the developing white matter have been described previously in depth (Hagemeyer et al., 2017; Wlodarczyk et al., 2017). This special white-matter microglia type plays crucial roles for the proper maintenance and maturation of oligodendrocyte progenitors and subsequent physiological myelinogenesis (Hagemeyer et al., 2017; Wlodarczyk et al., 2017). Moreover, this Clec7a+ subset of microglia seems to preferentially phagocytose cleaved caspase-3-positive dying newly formed oligodendrocytes during myelination (Li et al., 2019). The idea of microglial spatial heterogeneity already in the healthy CNS was further supported by a study using bulk RNAseq analysis that revealed regional heterogeneity of microglia that was more prominent in the cerebellum when compared to the cortex, hippocampus, or striatum (Grabert et al., 2016). Interestingly, recent scRNA-seq analyses that explored regional diversity failed to detect such a specificity of cerebellum microglia, but rather found a transcriptional continuum between microglia states across the CNS regions with few pre-existing states (Hammond et al., 2019; Keren-Shaul et al., 2017; Li et al., 2019; Masuda et al., 2019), highlighting a homogeneity of the microglia pool with no region-specific subclasses in adulthood. Li et al. (2019) claimed that certain genes that have been previously attributed to regional specification in cerebellum microglia were mainly expressed in non-microglia myeloid populations such as CAMs. Differential enrichment of distinct microglia clusters in the cerebellum was observed when compared to other regions, including the cortex and hippocampus, in another study (Masuda et al., 2019). Moreover, recent single-nucleus RNA-seq of striatal and cerebellar microglia from adult mice revealed epigenetic regulation for high-clearance activities in cerebellar microglia (Ayata et al., 2018), suggesting a functionally different state of microglia in the cerebellum. Further analysis will be needed to fully resolve this controversy. Another unexpected result by Li et al. (2019) was the unaltered transcriptome of microglia in the olfactory bulb, where microglia have been shown to exhibit a higher proliferation rate (Tay et al., 2017), and a distinct gene expression pattern in comparison with microglia from other CNS regions (Doorn et al., 2015). It remains unclear how many microglia from the olfactory bulb were analyzed in this study. One might consider that limited regional heterogeneity of homeostatic microglia observed in the recent studies might be due to too few microglial cells being analyzed or a too-shallow sequencing depth to further identify the transcriptionally distinct microglia states. Some possibilities can partially be excluded since the authors have taken advantage of well-established deep-sequencing protocols (e.g., Smartseq2) with a substantial number of microglial cells (Li et al., 2019; Masuda et al., 2019). Taken together, the consensus for adult homeostatic microglia so far is that there is limited spatial heterogeneity of microglia across anatomically distinct CNS regions, though further detailed analysis will be required to better understand regional diversity of microglia under homeostatic conditions. Gender and Microglia Heterogeneity Accumulating evidence has highlighted a striking impact of gender on gene expression profile and function of microglia during development and aging (Guneykaya et al., 2018;

Hanamsagar et al., 2017; Thion et al., 2018) and subsequent microglial inflammatory responses (Hanamsagar et al., 2017). Furthermore, the gender differences have been shown to affect the density, cell size, and phagocytic capacity of microglia (Schwarz et al., 2012; Weinhard et al., 2018). Hammond et al. (2019) have utilized scRNA-seq technologies to address the question of whether sexual dimorphisms in microglia affect their transcriptomes and diversity over the course of development. They analyzed transcripts of single microglia from males and females at different developmental stages (embryonic E14.5, postnatal P4/P5, and adult P100). Unexpectedly, unbiased clustering analysis did not segregate microglia in a sex-dependent manner but gave rise to only minor differences in cluster composition between the genders at the early postnatal time point (Hammond et al., 2019). These results indicate only subtle sex differences on microglia diversity in terms of gene expression during development and homeostasis. A follow-up analysis with a deeper sequencing protocol will help to confirm and better understand the impact of gender on microglia diversity. Microglia Diversity during Physiological Aging Microglia have been implicated in age-related neurodegenerative diseases such as Alzheimer’s disease and Parkinson’s disease (Krasemann et al., 2017). In addition, phenotypically reactive microglia have been shown to be present in the aged CNS, and their numbers increased with time (Flowers et al., 2017; Raj et al., 2017). A new scRNA-seq study provided further, deeper insights into microglia multiplicity during aging. Most microglia still clustered in the aged brains as normal homeostatic microglia, but aging obviously induced microglia to be more heterogeneous and especially induced immunogenic types with inflammatory and interferon-responsive profiles (Hammond et al., 2019; Figure 1). Among them, the inflammatory microglia subclass expressed a variety of inflammatory signaling molecules, including chemokines Ccl3 and Ccl4, Cst7 (encoding cystatin F), and Il1b (encoding interleukin 1 beta). The existence of an additional microglia state, characterized by a high expression of interferon-related genes such as interferon-induced transmembrane protein 3 (Ifitm3) and receptor transporter protein 4 (Rtp4), became prominent in the aged mouse brain as well (Hammond et al., 2019). Such age-related microglia have been implicated in the development of neuropathological conditions such as age-associated cognitive decline (Deczkowska et al., 2017). Because microglia are extremely plastic cells that readily change their phenotypes on demand, alterations occurring in the microenvironment may determine the features of such microglia states during senescence. Alternatively, it may be possible that microglia lose their intrinsic machinery that controls their proper function during aging, and such dysregulated microglia may be implicated in aging-related pathologies. Limited Microglia Heterogeneity after Depletion by CSF1 Receptor Inhibition Microglial viability is highly dependent on CSF1R signaling (Erblich et al., 2011) which recognizes CSF1 and alternatively interleukin (IL)-34 as ligands (Greter et al., 2012; Wang et al., 2012). Consequently, pharmacological inhibition of CSF1R leads to 99% ablation of microglia in the whole CNS (Huang et al.,

2018; Zhan et al., 2019). However, once the inhibitor is removed, the remaining microglia are able to rapidly proliferate and replenish the whole CNS within 1 week. A recent scRNA-seq analysis revealed a distinct microglial transcriptional profile during the repopulation phase, characterized by an upregulation of the cell-cycle-related genes Cdk1 and Mki67 and the migrationrelated gene Cd36 when compared to normal adult microglia (Huang et al., 2018). Interestingly, repopulating microglia were almost a homogeneous population with few distinct clusters, which suggests that regional cues of the CNS may not affect the proliferation potential of microglia; however, the authors did not compare the transcriptome of single repopulation microglia from different CNS regions at different repopulating time points (Huang et al., 2018). Considering that the dependency of microglial viability on the two biological ligands of CSF1R differs between the CNS regions (Greter et al., 2012; Wang et al., 2012), region- and time-specific microglial states might temporarily exist during repopulation after ablation. Context-Dependent Microglia Diversity during CNS Disease Progression In virtue of their plasticity, the homogeneity of homeostatic microglia is readily disrupted under pathological conditions. In general, a swift change in gene expression tailors ‘‘activated’’ microglia to react on demand to surrounding alterations. Keren-Shaul et al. (2017) identified microglial clusters with distinct gene expressional profiles in the cortical regions in an animal model for Alzheimer’s disease where Ab plaques are massively deposited (Figure 2). As mentioned above, they referred to these microglia as DAM. DAM presented a notable gene signature with high expression levels of Lpl (encoding lipoprotein lipase), Itgax (encoding CD11c), Apoe, and Cst7 (Keren-Shaul et al., 2017). Histological analysis confirmed the localization of these microglia subtypes to the area surrounding Ab plaques. scRNA-seq, analysis together with computational clustering, grouped DAM into two different stages (DAM stages 1 and 2), indicating more diverse microglia states during disease progression. The authors further elegantly defined the molecular machinery for the determination of the DAM phenotype. The transition from the normal state to stage 1 of DAM is TREM2 independent, whereas shifting further from stage 1 to 2 requires TREM2 signals (Keren-Shaul et al., 2017), suggesting complex mechanisms facilitating microglial diversity. DAM were also detected in mice carrying mutated SOD1, a mouse model of familial amyotrophic lateral sclerosis (ALS) (Tu et al., 1997). Peripheral lipopolysaccharide (LPS) injection has often been used to mimic systemic inflammation and sterile infection, which induces alterations in microglia via complex mechanisms (Jeong et al., 2010; Serrats et al., 2010). A recent study adapted this system to explore the heterogeneity of the microglial response to the inflammation at single-cell resolution (Sousa et al., 2018). As expected, almost all microglia in the LPS-injected mice changed their gene expression profiles, as shown previously with bulk RNA-seq analysis (Friedman et al., 2018). However, after LPS treatment, microglia showed a unique gene expression profile that was different from that of DAM. The expression of genes such as Trem2, TYRO protein tyrosine kinase-binding protein (Tyrobp), Ctsd, or hypoxia-inducible factor 1-alpha (Hif1a) was

Cell Reports 30, February 4, 2020 1275

Figure 2. Profiles of Context-Dependent Microglial Clusters during Defined Disease Conditions in Mice Microglia phenotypes with distinct gene expressional profiles appear under disease conditions in a disease-context-dependent fashion. Key transcriptional profiles are shown. EAE, experimental autoimmune encephalomyelitis; 5xFAD, five familial Alzheimer’s disease mutations; FNA, facial nerve axotomy; MG, disease-associated microglia; mSOD1, mutant human superoxide dismutase 1.

upregulated in DAM but downregulated in LPS-associated microglia, suggesting different types of reactive microglia states (Sousa et al., 2018). These findings indicate that the appearance of unique microglial subclasses is related to the inflammatory environment, which can be phenotypically distinct from DAM. A recent study addressed the important question of diseasespecific microglia clusters by comparing microglial activation states induced either by axotomizing the facial nerve (FNX), which causes subsequent neuronal degeneration, or by administering a copper chelating toxin cuprizone, which causes cell death of oligodendrocytes by disrupting metabolic processes (Masuda et al., 2019). Both FNX and cuprizone treatment can cause local microglia activation without any cell infiltration from the circulation (Mildner et al., 2007; Tay et al., 2017). Indeed, in both disease models, microglia were activated and changed their gene expression patterns, whereas the induced phenotypes were clearly transcriptionally separated (Masuda et al., 2019). Microglia from cuprizone-fed mice were even subdivided into two subclusters with distinct gene expression profiles. One subset was observed mainly during the demyelinating stage, and the other was related to the remyelination. Notably, microglia derived from FNX exhibited a strongly increased expression of genes including Ctsc (encoding cathepsin C), whereas demyelination-related microglia phenocopied that of DAM by upregulating a variety of genes such as Spp1 or Cst7. Spp1 is known to have the potential to enhance myelination (Selvaraju et al., 2004). In addition, the knockdown of the Cst7 gene has been

1276 Cell Reports 30, February 4, 2020

shown to aggravate the demyelination in the cuprizone model (Liang et al., 2016), suggesting the regulatory role of this microglia subset during demyelination. In contrast, the remyelination-related subset of microglia was characterized by a high expression of Cybb (encoding NADPH oxidase 2) and major histocompatibility complex (MHC) class-II-related genes such as Cd74 and H2-Aa, demonstrating the existence of a variety of DAM subtypes that depend on disease contexts (e.g., neurodegenerative conditions versus toxic demyelination). The precise functions of respective microglia clusters during disease remain to be elucidated. While such de- or re-myelination-associated microglia shared transcriptional features to some extent with early embryonic microglia that were analyzed in parallel, those microglia sets were still distinctly different (Masuda et al., 2019). This suggests that complex molecular machinery controls the characteristics of a variety of microglia states in a contextdependent manner, giving rise to highly specialized clusters of microglia with distinct functions. Prominent heterogeneity of microglia was also observed in the CNS of mice afflicted with experimental autoimmune encephalomyelitis (EAE) as another demyelination model. During this type of sterile neuroinflammation, microglia exhibited greater diversity and gave rise to several disease-associated ~o phenotypes with distinct gene expression patterns (Jorda et al., 2019). The EAE-associated microglia cluster massively decreased the expression of typical microglia core genes such as P2ry12, Tmem119, and Selplg (encoding P-selectin glycoprotein ligand-1) and upregulated the expression of Ly86, Ccl2, Cxcl10, and Mki67 (encoding marker of proliferation Ki-67), proposing a high proliferative capacity of these microglia states in mice. Each distinct DAM cluster was separated by different levels of Cxcl10, Cd74, and Ccl5 transcripts, and the three respective subsets (CXCL10lo CD74hi, CXCL10hi CD74lo, or CCL5+CD74lo) could be confirmed histologically on the protein ~o et al., 2019). Similarly, Hammond et al. (2019) inlevel (Jorda jected mice with lysolecithin (LPC), a different model of focal demyelination, and also identified heterogeneous microglia states that were characterized by high expression of Ccl4, Cxcl10, or Birc5 (encoding baculoviral inhibitor of apoptosis repeat-containing 5). Although the function of each microglia phenotype is still unknown, given the fact that some chemokines—including CXCL10 and CCL5—are known to be chemotactic for peripheral leukocytes (e.g., T cells), such microglia clusters may play a role in recruiting leukocytes, which causes inflammation in the affected tissue. Of note, despite the similarities among three different disease models with demyelination (cuprizone, EAE, and LPC), the characteristics of defined DAM in each model differed. This may be explained by context dependency such as distinct cellular composition (e.g., massive infiltration of peripherally derived cells into CNS parenchyma takes place during EAE, whereas cuprizone treatment causes local microglial activation without the recruitment of circulating monocytes; Mildner et al., 2009). Furthermore, the use of different CNS regions (corpus callosum versus spinal cord) or the different background of mice used (CD1 versus C57BL/6) might also have an impact on microglia features. Nevertheless, all these independent studies led to the same conclusion: disease-related perturbation causes a swift

Figure 3. Diversity of Human Microglia under Homeostatic and Diseased Conditions Recent studies with single-cell analysis identified transcriptional heterogeneity of human microglia under both homeostatic and disease conditions. Greater diversity of microglia was found in the brains of patients with multiple sclerosis (MS). On the other hand, a large-scale analysis of human microglia with CyTOF identified regional heterogeneity in normal brains.

change in microglial gene expression, which is essentially shaped by the underlying pathology leading to higher microglial diversity. Although there are clearly disease-condition-specific microglial activation states, an important question is not yet answered: What is the ‘‘context’’ that controls the state and subsets of microglia in the respective disease conditions? It is tempting to speculate that both intrinsic (e.g., transcription factors) and extrinsic (e.g., resolution factors) signals might be involved, but the precise nature of such factors is still unknown. The identification of those elements would be critically important since the appearance of context-dependent microglia subtypes with their own specific transcriptional profiles has potential therapeutic implications. In sum, exciting new data from recent studies in mice indicated a surprisingly high transcriptional homogeneity of microglia with few pre-existing subclasses under homeostatic conditions, whereas microglia subtype specification emerges during challenges in a region- and disease-stage-specific manner (Figures 1 and 2). Microglial Multiplicity in Human Brains Devoid of Apparent Pathology With the rapid development of experimental tools, human microglia have been studied recently in order to profile their gene expression and to translate the well-established mouse data to the human context (Galatro et al., 2017; Zhong et al., 2018). Analysis of orthologous human and mouse transcriptomics has

demonstrated a high similarity of microglia between the two species (Gosselin et al., 2017). Of note, there are only a limited number of studies that profiled the transcriptomes of human microglia on the single-cell level, and human microglia heterogeneity on a spatial or temporal axis has only recently been addressed. The scarcity of studies is largely due to the objective difficulties in obtaining ‘‘healthy’’, non-damaged, freshly isolated human microglia. Thus, most of the recent studies used neurosurgical access material that was obtained during neurooncological or epileptic surgery. Indeed, these microscopically ‘‘normal’’ CNS specimens can still be affected to some extent by the underlying pathology. Novel studies addressed this open question by profiling human microglia landscapes at single-cell resolution (Figure 3). A first scRNA-seq analysis with freshly isolated microglia cells from surgically resected, non-affected regions of epileptic patient brains was performed just recently (Masuda et al., 2019). The authors identified distinct homeostatic human microglia states with some similarity to the gene expression profiles (e.g., CST3 expression) of mouse homeostatic microglia (Masuda et al., 2019). By contrast, differences between the two species were highlighted with small microglia substates present only in humans, which showed a comparatively high expression of chemokine genes (e.g., CCL4 and CCL2) and zinc finger transcription factors (e.g., EGR2 and EGR3), indicating a pre-active state (Masuda et al., 2019). This subset may reflect inflammatory changes in the microenvironment of the patient brains or represent a physiological human-specific microglia state, as suggested before (Darmanis et al., 2015). Overall, there was clearly greater microglia diversity in the human brains, detectable with only partially overlapping gene expression profiles, when compared to mouse microglia. However, comparisons of microglia from different brain regions of healthy patients were not conducted in this study. Another recent study using a combination of several singlecell methods—namely, scRNA-seq and multiplexed mass cytometry (CyTOF)—identified several microglia states with distinct gene expression profiles and demonstrated spatial diversity by comparing microglia from temporal lobe gray matter and white matter (Sankowski et al., 2019). Among the identified microglia subclusters, microglia characterized by higher expression levels of the MHC class-II factor HLA-DR and CD68 were enriched in white matter, whereas gray-matter microglia expressed those genes at a lower level, demonstrating spatial heterogeneity of microglia also in humans. The authors additionally demonstrated the age-dependent existence of a SPP1+ microglial subclass, which was more prominent in the aged brains (>50 years) compared to younger brains (Sankowski et al., 2019). Although the role of respective microglial types is still unknown, these data highlight higher spatial and temporal diversity of microglia in human brains when compared to microglia from laboratory mice. Comprehensive analysis of human microglia with CyTOF with barcoding technologies recently provided reliable information about regional heterogeneity by analyzing the landscape of human microglia (Bo¨ttcher et al., 2019). In this study, the immune phenotype of single microglia isolated from up to five different brain regions of post-mortem cryopreserved human brains—

Cell Reports 30, February 4, 2020 1277

including the subventricular zone, thalamus, cerebellum, temporal lobe, and frontal lobe—was deeply profiled (Bo¨ttcher et al., 2019). Interestingly, human microglia constituted at least four defined subtypes, with different enrichment of each subset across the regions tested. For example, one microglia subcluster enriched in the subventricular zone and thalamus showed a high expression of distinct molecules such as CD68, CD11c, CD45, CD64, CD195, HLA-DR, and the proliferation markers cyclin A and cyclin B1, indicating an active state of microglia in these regions. In contrast, the other two microglia clusters expressing higher levels of CD206 were more abundant in the temporal lobe and frontal lobe. Although there was no region-specific microglia subset in the cerebellum, the protein expression profile of cerebellar microglia differed from that of the other regions (Bo¨ttcher et al., 2019). Whether these subclusters of human microglia enriched in respective brain regions reflect a region-specific function of microglia remains unknown. If this is the case, determining the molecular machinery that induces regional diversity of microglia will help to better understand the nature of homeostatic microglia in humans. Together, human brains seem to comprise more diverse states of microglia with distinct gene expression profiles in a region-specific manner than those found in mice. Further detailed analysis will be required to examine the mechanisms by which such diversity in human microglia is induced. Microglial States in Diseased Human Brains A pathology-dependent environment in the CNS has been shown to induce a specific transcriptional network that specifies microglial identity in humans (Gosselin et al., 2017). Defining the existence of microglial clusters in humans during disease is ultimately important to understand the roles of the respective microglia types for disease pathogenesis and to provide new potential therapeutic targets (Priller and Prinz, 2019). As mentioned, accumulating evidence has shown the existence of DAM in animal disease models, whereas much less is known about such disease-specific subtypes in human disease. A new study tackled this point by performing scRNA-seq analysis with microglia isolated from the brains of patients with multiple sclerosis (MS) (Masuda et al., 2019), a debilitating neurological disease that is characterized by an autoimmune demyelination of the CNS (Voet et al., 2019). Analyses of single microglia transcriptomes from MS patients and healthy controls identified several microglia phenotypes (e.g., three common microglia clusters that expressed homeostatic microglia core genes such as TMEM119, P2RY12, and SLC2A5 [encoding solute carrier family 2 member 5] and one common cluster that showed lower expression of microglia core genes and high expression of chemokines [CCL2 and CCL4)]). The authors further identified one MS-associated and two MS-enriched groups of microglia that downregulated microglia core genes and increased levels of expression of APOE and MAFB (encoding V-maf musculoaponeurotic fibrosarcoma oncogene homolog B) (Masuda et al., 2019). The MSassociated microglia cluster was characterized by a high expression of CTSD, APOC1 (encoding apolipoprotein C1), and GPNMB. The one MS-enriched cluster showed increased expression of MHC class-II-related molecules such as CD74, HLA-DRA, and HLA-DRB1, whereas strong expression of

1278 Cell Reports 30, February 4, 2020

SPP1, PADI2, and LPL transcripts was seen in the other MS-enriched populations. Of note, the proportion of each state in the brains of individual patients varied greatly, indicating high interindividual heterogeneity of microglia during the disease progression of MS. Furthermore, canonical correlation analysis of mouse and human microglia orthologs confirmed the similarity of transcriptomic profiles between these MS-related microglia phenotypes and mouse microglia clusters observed in the demyelinating disease model (Masuda et al., 2019; Figure 3). A further recent study explored human microglial heterogeneity and the appearance of disease-associated states by profiling microglia from glioblastomas by combining scRNA-seq and CyTOF (Sankowski et al., 2019). In general, glioma-associated microglia (GAM) downregulated microglia core genes including CX3CR1 and SELPLG, whereas several distinct groups of GAM were separated by different expression levels of hypoxia-associated genes such as HIF1A and VEGFA (encoding vascular endothelial growth factor A), the interferon-related gene IFI44 (encoding interferon induced protein 44), SPP1, HLA-DRA, APOE, or CD163 (Sankowski et al., 2019), demonstrating highly heterogeneous microglia in human gliomas. The functional relevance of microglia in the glioma remains highly speculative, ranging from being a mere bystander to an active promoter for glioma growth (Hambardzumyan et al., 2016). Among the identified subclasses, VEGFAhi microglia and CD163hi microglia were identified as context-dependent subclusters in gliomas (Sankowski et al., 2019), which may exert a highly specialized function for glioma growth. Understanding the nature of such microglia clusters will help to better understand the biology of gliomas and to develop new therapeutics (Priller and Prinz, 2019). Although the context dependency of microglia phenotypes in the diseased human CNS has not yet been examined, the further characterization of microglia subclasses in humans would have an enormous impact and provide deeper insights into the pathogenesis of CNS disorders. Open Questions and Future Directions The recent advance for single-cell analysis shed exciting new light on the cellular diversity in the whole body, including the CNS. Here, we have outlined recent findings that uncovered the heterogeneity of microglia in both mice and humans on a temporal and spatial axis that give novel insights into the nature of microglia during development, homeostasis, aging, and disease progression. Nevertheless, there is still a long list of open questions that need to be answered. While recent single-cell studies have elegantly identified previously unknown clusters of microglia during both development and perturbation, the distinct roles of these respective states remain elusive. Application of selective tools targeting only the context-dependent microglia cluster by Cre recombinase-mediated gene targeting or defined drugs might provide new insights into the nature of microglia in the CNS. Technically, transcriptomic or proteomic measurements of cells cannot capture the epigenetic heterogeneity that may drive cellular biology. In addition, differentially regulated proteins, which could not be captured by CyTOF analysis with a limited number of target molecules, may ensure functional diversity in microglia. Therefore, results from scRNA-seq or CyTOF

experiments may describe only a subset of microglia molecular phenotypes. A recent study proposed regional differences in deep-brain microglia such as those within the basal ganglia, where microglia were found to differ in morphology, density, and membrane properties (De Biase et al., 2017). All scRNA-seq studies to date did not analyze microglia from these regions, and further analysis is required to determine whether the observed differences in the membrane properties of microglia groups are functionally relevant. Future studies covering comprehensive analyses with functional readouts will give deeper insights into the microglial heterogeneity on a functional basis. In addition, other technologies for single-cell analysis, such as single-cell ATAC-sequencing or similar methods to study chromatin accessibility, or single-cell protein analysis (e.g., CITE-seq, REAP-seq) will help to achieve comprehensive analysis. Recent studies have highlighted the importance of host microbiota for shaping the functional phenotype of microglia under both homeostatic and pathological conditions (Erny et al., 2015; Matcovitch-Natan et al., 2016; Thion et al., 2018). It will be important in the future to reveal how, and to what extent, the host microbiota affect microglial diversity in the CNS during normal development as well as during disease progression. To date, little is known about the temporal heterogeneity of human microglia on the transcriptional level during aging, where microglia seem to show a distinct gene signature that is linked to the altered microenvironment of the CNS (Galatro et al., 2017). Analyzing microglia from aged human brains will illuminate another aspect of microglial diversity in humans and help to identify specific subsets with specialized functions for the pathogenesis of age-related diseases. In addition, whether such microglia phenotypes are detrimental or beneficial for disease progression in both mice and humans still remains unclear. Answering such critical questions together with profiling heterogeneous subclusters of microglia in healthy and diseased situations can open new avenues for the development of therapeutic targeting of microglia. ACKNOWLEDGMENTS

REFERENCES Abiega, O., Beccari, S., Diaz-Aparicio, I., Nadjar, A., Laye´, S., Leyrolle, Q., Go´mez-Nicola, D., Domercq, M., Pe´rez-Samartı´n, A., Sa´nchez-Zafra, V., et al. (2016). Neuronal Hyperactivity Disturbs ATP Microgradients, Impairs Microglial Motility, and Reduces Phagocytic Receptor Expression Triggering Apoptosis/Microglial Phagocytosis Uncoupling. PLoS Biol. 14, e1002466. Ajami, B., Samusik, N., Wieghofer, P., Ho, P.P., Crotti, A., Bjornson, Z., Prinz, M., Fantl, W.J., Nolan, G.P., and Steinman, L. (2018). Single-cell mass cytometry reveals distinct populations of brain myeloid cells in mouse neuroinflammation and neurodegeneration models. Nat. Neurosci. 21, 541–551. Askew, K., Li, K., Olmos-Alonso, A., Garcia-Moreno, F., Liang, Y., Richardson, P., Tipton, T., Chapman, M.A., Riecken, K., Beccari, S., et al. (2017). Coupled Proliferation and Apoptosis Maintain the Rapid Turnover of Microglia in the Adult Brain. Cell Rep. 18, 391–405. Ayata, P., Badimon, A., Strasburger, H.J., Duff, M.K., Montgomery, S.E., Loh, Y.E., Ebert, A., Pimenova, A.A., Ramirez, B.R., Chan, A.T., et al. (2018). Epigenetic regulation of brain region-specific microglia clearance activity. Nat. Neurosci. 21, 1049–1060. Bo¨ttcher, C., Schlickeiser, S., Sneeboer, M.A.M., Kunkel, D., Knop, A., Paza, E., Fidzinski, P., Kraus, L., Snijders, G.J.L., Kahn, R.S., et al.; NBB-Psy (2019). Human microglia regional heterogeneity and phenotypes determined by multiplexed single-cell mass cytometry. Nat. Neurosci. 22, 78–90. Bruttger, J., Karram, K., Wo¨rtge, S., Regen, T., Marini, F., Hoppmann, N., Klein, M., Blank, T., Yona, S., Wolf, Y., et al. (2015). Genetic Cell Ablation Reveals Clusters of Local Self-Renewing Microglia in the Mammalian Central Nervous System. Immunity 43, 92–106. Butovsky, O., and Weiner, H.L. (2018). Microglial signatures and their role in health and disease. Nat. Rev. Neurosci. 19, 622–635. Colonna, M., and Butovsky, O. (2017). Microglia Function in the Central Nervous System During Health and Neurodegeneration. Annu. Rev. Immunol. 35, 441–468. Darmanis, S., Sloan, S.A., Zhang, Y., Enge, M., Caneda, C., Shuer, L.M., Hayden Gephart, M.G., Barres, B.A., and Quake, S.R. (2015). A survey of human brain transcriptome diversity at the single cell level. Proc. Natl. Acad. Sci. USA 112, 7285–7290. De Biase, L.M., Schuebel, K.E., Fusfeld, Z.H., Jair, K., Hawes, I.A., Cimbro, R., Zhang, H.Y., Liu, Q.R., Shen, H., Xi, Z.X., et al. (2017). Local cues establish and maintain region-specific phenotypes of basal ganglia microglia. Neuron 95, 341–356.e346. Deczkowska, A., Matcovitch-Natan, O., Tsitsou-Kampeli, A., Ben-Hamo, S., Dvir-Szternfeld, R., Spinrad, A., Singer, O., David, E., Winter, D.R., Smith, L.K., et al. (2017). Mef2C restrains microglial inflammatory response and is lost in brain ageing in an IFN-I-dependent manner. Nat. Commun. 8, 717.

T.M. was supported by the KANAE Foundation for the Promotion of Medical Science and the Japan Society for the Promotion of Science (JSPS) as the JSPS Postdoctoral Fellow for Research Abroad. M.P. is supported by the Sobek Foundation, the Ernst-Jung Foundation, the German Research Foundation (SFB 992, SFB1160, Reinhart-Koselleck grant), the Ministry of Science, Research and Arts, Baden-Wuerttemberg (Sonderlinie ‘‘Neuroinflammation’’), and the BMBF-funded competence network of multiple sclerosis (KKNMS). This study was supported by the German Research Foundation (DFG) under Germany’s Excellence Strategy (CIBSS – EXC-2189 – project ID 390939984).

Dissing-Olesen, L., LeDue, J.M., Rungta, R.L., Hefendehl, J.K., Choi, H.B., and MacVicar, B.A. (2014). Activation of neuronal NMDA receptors triggers transient ATP-mediated microglial process outgrowth. J. Neurosci. 34, 10511– 10527.

AUTHOR CONTRIBUTIONS

 de Angelis, A.L., Jaitin, D., Wieghofer, P., Staszewski, O., DaErny, D., Hrabe vid, E., Keren-Shaul, H., Mahlakoiv, T., Jakobshagen, K., Buch, T., et al. (2015). Host microbiota constantly control maturation and function of microglia in the CNS. Nat. Neurosci. 18, 965–977.

All authors contributed to researching the data for the article and writing the article. T.M. and M.P. made substantial contributions to the discussion of the content of the article.

Doorn, K.J., Breve´, J.J., Drukarch, B., Boddeke, H.W., Huitinga, I., Lucassen, P.J., and van Dam, A.M. (2015). Brain region-specific gene expression profiles in freshly isolated rat microglia. Front. Cell. Neurosci. 9, 84. Erblich, B., Zhu, L., Etgen, A.M., Dobrenis, K., and Pollard, J.W. (2011). Absence of colony stimulation factor-1 receptor results in loss of microglia, disrupted brain development and olfactory deficits. PLoS ONE 6, e26317.

DECLARATION OF INTERESTS

Flowers, A., Bell-Temin, H., Jalloh, A., Stevens, S.M., Jr., and Bickford, P.C. (2017). Proteomic anaysis of aged microglia: shifts in transcription, bioenergetics, and nutrient response. J. Neuroinflammation 14, 96.

The authors declare no competing interests.

Fontainhas, A.M., Wang, M., Liang, K.J., Chen, S., Mettu, P., Damani, M., Fariss, R.N., Li, W., and Wong, W.T. (2011). Microglial morphology and dynamic

Cell Reports 30, February 4, 2020 1279

behavior is regulated by ionotropic glutamatergic and GABAergic neurotransmission. PLoS ONE 6, e15973.

proliferation of residual microglia after acute depletion. Nat. Neurosci. 21, 530–540.

Friedman, B.A., Srinivasan, K., Ayalon, G., Meilandt, W.J., Lin, H., Huntley, M.A., Cao, Y., Lee, S.H., Haddick, P.C.G., Ngu, H., et al. (2018). Diverse Brain Myeloid Expression Profiles Reveal Distinct Microglial Activation States and Aspects of Alzheimer’s Disease Not Evident in Mouse Models. Cell Rep. 22, 832–847.

Jeong, H.K., Jou, I., and Joe, E.H. (2010). Systemic LPS administration induces brain inflammation but not dopaminergic neuronal death in the substantia nigra. Exp. Mol. Med. 42, 823–832.

Galatro, T.F., Holtman, I.R., Lerario, A.M., Vainchtein, I.D., Brouwer, N., Sola, P.R., Veras, M.M., Pereira, T.F., Leite, R.E.P., Mo¨ller, T., et al. (2017). Transcriptomic analysis of purified human cortical microglia reveals age-associated changes. Nat. Neurosci. 20, 1162–1171. Ginhoux, F., Greter, M., Leboeuf, M., Nandi, S., See, P., Gokhan, S., Mehler, M.F., Conway, S.J., Ng, L.G., Stanley, E.R., et al. (2010). Fate mapping analysis reveals that adult microglia derive from primitive macrophages. Science 330, 841–845. Goldmann, T., Zeller, N., Raasch, J., Kierdorf, K., Frenzel, K., Ketscher, L., Basters, A., Staszewski, O., Brendecke, S.M., Spiess, A., et al. (2015). USP18 lack in microglia causes destructive interferonopathy of the mouse brain. EMBO J. 34, 1612–1629. ~o, M.J., Prutek, F., Hagemeyer, N., Frenzel, Goldmann, T., Wieghofer, P., Jorda K., Amann, L., Staszewski, O., Kierdorf, K., Krueger, M., et al. (2016). Origin, fate and dynamics of macrophages at central nervous system interfaces. Nat. Immunol. 17, 797–805. Gomez Perdiguero, E., Klapproth, K., Schulz, C., Busch, K., Azzoni, E., Crozet, L., Garner, H., Trouillet, C., de Bruijn, M.F., Geissmann, F., and Rodewald, H.R. (2015). Tissue-resident macrophages originate from yolk-sac-derived erythromyeloid progenitors. Nature 518, 547–551.

~o, M.J.C., Sankowski, R., Brendecke, S.M., Sagar, Locatelli, G., Tai, Jorda Y.H., Tay, T.L., Schramm, E., Armbruster, S., Hagemeyer, N., et al. (2019). Single-cell profiling identifies myeloid cell subsets with distinct fates during neuroinflammation. Science 363, eaat7554. Keren-Shaul, H., Spinrad, A., Weiner, A., Matcovitch-Natan, O., Dvir-Szternfeld, R., Ulland, T.K., David, E., Baruch, K., Lara-Astaiso, D., Toth, B., et al. (2017). A Unique Microglia Type Associated with Restricting Development of Alzheimer’s Disease. Cell 169, 1276–1290.e1217. Kierdorf, K., Erny, D., Goldmann, T., Sander, V., Schulz, C., Perdiguero, E.G., Wieghofer, P., Heinrich, A., Riemke, P., Ho¨lscher, C., et al. (2013). Microglia emerge from erythromyeloid precursors via Pu.1- and Irf8-dependent pathways. Nat. Neurosci. 16, 273–280. ~o, M.J.C., and Prinz, M. (2019). Macrophages at Kierdorf, K., Masuda, T., Jorda CNS interfaces: ontogeny and function in health and disease. Nat. Rev. Neurosci. 20, 547–562. Krasemann, S., Madore, C., Cialic, R., Baufeld, C., Calcagno, N., El Fatimy, R., Beckers, L., O’Loughlin, E., Xu, Y., Fanek, Z., et al. (2017). The TREM2-APOE Pathway Drives the Transcriptional Phenotype of Dysfunctional Microglia in Neurodegenerative Diseases. Immunity 47, 566–581.e569. Lawson, L.J., Perry, V.H., Dri, P., and Gordon, S. (1990). Heterogeneity in the distribution and morphology of microglia in the normal adult mouse brain. Neuroscience 39, 151–170.

Gosselin, D., Skola, D., Coufal, N.G., Holtman, I.R., Schlachetzki, J.C.M., Sajti, E., Jaeger, B.N., O’Connor, C., Fitzpatrick, C., Pasillas, M.P., et al. (2017). An environment-dependent transcriptional network specifies human microglia identity. Science 356, eaal3222.

Li, Q., Cheng, Z., Zhou, L., Darmanis, S., Neff, N.F., Okamoto, J., Gulati, G., Bennett, M.L., Sun, L.O., Clarke, L.E., et al. (2019). Developmental Heterogeneity of Microglia and Brain Myeloid Cells Revealed by Deep Single-Cell RNA Sequencing. Neuron 101, 207–223.e210.

Grabert, K., Michoel, T., Karavolos, M.H., Clohisey, S., Baillie, J.K., Stevens, M.P., Freeman, T.C., Summers, K.M., and McColl, B.W. (2016). Microglial brain region-dependent diversity and selective regional sensitivities to aging. Nat. Neurosci. 19, 504–516.

Liang, J., Li, N., Zhang, Y., Hou, C., Yang, X., Shimizu, T., Wang, X., Ikenaka, K., Fan, K., and Ma, J. (2016). Disinhibition of Cathepsin C Caused by Cystatin F Deficiency Aggravates the Demyelination in a Cuprizone Model. Front. Mol. Neurosci. 9, 152.

€ndig, Greter, M., Lelios, I., Pelczar, P., Hoeffel, G., Price, J., Leboeuf, M., Ku T.M., Frei, K., Ginhoux, F., Merad, M., and Becher, B. (2012). Stroma-derived interleukin-34 controls the development and maintenance of langerhans cells and the maintenance of microglia. Immunity 37, 1050–1060.

Masuda, T., and Prinz, M. (2016). Microglia: A Unique Versatile Cell in the Central Nervous System. ACS Chem. Neurosci. 7, 428–434.

Guneykaya, D., Ivanov, A., Hernandez, D.P., Haage, V., Wojtas, B., Meyer, N., Maricos, M., Jordan, P., Buonfiglioli, A., Gielniewski, B., et al. (2018). Transcriptional and translational differences of microglia from male and female brains. Cell Rep. 24, 2773–2783.e2776. Hagemeyer, N., Hanft, K.M., Akriditou, M.A., Unger, N., Park, E.S., Stanley, E.R., Staszewski, O., Dimou, L., and Prinz, M. (2017). Microglia contribute to normal myelinogenesis and to oligodendrocyte progenitor maintenance during adulthood. Acta Neuropathol. 134, 441–458. Hambardzumyan, D., Gutmann, D.H., and Kettenmann, H. (2016). The role of microglia and macrophages in glioma maintenance and progression. Nat. Neurosci. 19, 20–27. Hammond, T.R., Dufort, C., Dissing-Olesen, L., Giera, S., Young, A., Wysoker, A., Walker, A.J., Gergits, F., Segel, M., Nemesh, J., et al. (2019). Single-Cell RNA Sequencing of Microglia throughout the Mouse Lifespan and in the Injured Brain Reveals Complex Cell-State Changes. Immunity 50, 253– 271.e256. Hanamsagar, R., Alter, M.D., Block, C.S., Sullivan, H., Bolton, J.L., and Bilbo, S.D. (2017). Generation of a microglial developmental index in mice and in humans reveals a sex difference in maturation and immune reactivity. Glia 65, 1504–1520.

Masuda, T., Sankowski, R., Staszewski, O., Bo¨ttcher, C., Amann, L., Sagar, Scheiwe, C., Nessler, S., Kunz, P., van Loo, G., et al. (2019). Spatial and temporal heterogeneity of mouse and human microglia at single-cell resolution. Nature 566, 388–392. Matcovitch-Natan, O., Winter, D.R., Giladi, A., Vargas Aguilar, S., Spinrad, A., Sarrazin, S., Ben-Yehuda, H., David, E., Zelada Gonza´lez, F., Perrin, P., et al. (2016). Microglia development follows a stepwise program to regulate brain homeostasis. Science 353, aad8670. Mathys, H., Adaikkan, C., Gao, F., Young, J.Z., Manet, E., Hemberg, M., De Jager, P.L., Ransohoff, R.M., Regev, A., and Tsai, L.-H. (2017). Temporal Tracking of Microglia Activation in Neurodegeneration at Single-Cell Resolution. Cell Rep. 21, 366–380. Mildner, A., Schmidt, H., Nitsche, M., Merkler, D., Hanisch, U.K., Mack, M., €ck, W., Priller, J., and Prinz, M. (2007). Microglia in the Heikenwalder, M., Bru adult brain arise from Ly-6ChiCCR2+ monocytes only under defined host conditions. Nat. Neurosci. 10, 1544–1553. €ck, W., Djukic, M., Zabel, M.D., Hille, A., Mildner, A., Mack, M., Schmidt, H., Bru Priller, J., and Prinz, M. (2009). CCR2+Ly-6Chi monocytes are crucial for the effector phase of autoimmunity in the central nervous system. Brain 132, 2487–2500.

Hanisch, U.K. (2013). Functional diversity of microglia - how heterogeneous are they to begin with? Front. Cell. Neurosci. 7, 65.

Mildner, A., Huang, H., Radke, J., Stenzel, W., and Priller, J. (2017). P2Y12 receptor is expressed on human microglia under physiological conditions throughout development and is sensitive to neuroinflammatory diseases. Glia 65, 375–387.

Huang, Y., Xu, Z., Xiong, S., Sun, F., Qin, G., Hu, G., Wang, J., Zhao, L., Liang, Y.X., Wu, T., et al. (2018). Repopulated microglia are solely derived from the

Mrdjen, D., Pavlovic, A., Hartmann, F.J., Schreiner, B., Utz, S.G., Leung, B.P., Lelios, I., Heppner, F.L., Kipnis, J., Merkler, D., et al. (2018). High-Dimensional

1280 Cell Reports 30, February 4, 2020

Single-Cell Mapping of Central Nervous System Immune Cells Reveals Distinct Myeloid Subsets in Health, Aging, and Disease. Immunity 48, 380– 395.e386. Priller, J., and Prinz, M. (2019). Targeting microglia in brain disorders. Science 365, 32–33. Prinz, M., and Priller, J. (2014). Microglia and brain macrophages in the molecular age: from origin to neuropsychiatric disease. Nat. Rev. Neurosci. 15, 300–312. Prinz, M., Erny, D., and Hagemeyer, N. (2017). Ontogeny and homeostasis of CNS myeloid cells. Nat. Immunol. 18, 385–392. Prinz, M., Jung, S., and Priller, J. (2019). Microglia Biology: One Century of Evolving Concepts. Cell 179, 292–311. Raj, D., Yin, Z., Breur, M., Doorduin, J., Holtman, I.R., Olah, M., MantinghOtter, I.J., Van Dam, D., De Deyn, P.P., den Dunnen, W., et al. (2017). Increased White Matter Inflammation in Aging- and Alzheimer’s Disease Brain. Front. Mol. Neurosci. 10, 206. Ransohoff, R.M., and Perry, V.H. (2009). Microglial physiology: unique stimuli, specialized responses. Annu. Rev. Immunol. 27, 119–145. Ren, L., Lubrich, B., Biber, K., and Gebicke-Haerter, P.J. (1999). Differential expression of inflammatory mediators in rat microglia cultured from different brain regions. Brain Res. Mol. Brain Res. 65, 198–205. Rothhammer, V., Borucki, D.M., Tjon, E.C., Takenaka, M.C., Chao, C.C., Ardura-Fabregat, A., de Lima, K.A., Gutie´rrez-Va´zquez, C., Hewson, P., Staszewski, O., et al. (2018). Microglial control of astrocytes in response to microbial metabolites. Nature 557, 724–728. Sankowski, R., Bo¨ttcher, C., Masuda, T., Geirsdottir, L., Sagar, Sindram, E., Seredenina, T., Muhs, A., Scheiwe, C., Shah, M.J., et al. (2019). Mapping microglia states in the human brain through the integration of high-dimensional techniques. Nat. Neurosci. 22, 2098–2110. Schmid, C.D., Sautkulis, L.N., Danielson, P.E., Cooper, J., Hasel, K.W., Hilbush, B.S., Sutcliffe, J.G., and Carson, M.J. (2002). Heterogeneous expression of the triggering receptor expressed on myeloid cells-2 on adult murine microglia. J. Neurochem. 83, 1309–1320. Schulz, C., Gomez Perdiguero, E., Chorro, L., Szabo-Rogers, H., Cagnard, N., Kierdorf, K., Prinz, M., Wu, B., Jacobsen, S.E., Pollard, J.W., et al. (2012). A lineage of myeloid cells independent of Myb and hematopoietic stem cells. Science 336, 86–90. Schwarz, J.M., Sholar, P.W., and Bilbo, S.D. (2012). Sex differences in microglial colonization of the developing rat brain. J. Neurochem. 120, 948–963. Selvaraju, R., Bernasconi, L., Losberger, C., Graber, P., Kadi, L., AvellanaAdalid, V., Picard-Riera, N., Baron Van Evercooren, A., Cirillo, R., Kosco-Vilbois, M., et al. (2004). Osteopontin is upregulated during in vivo demyelination and remyelination and enhances myelin formation in vitro. Mol. Cell. Neurosci. 25, 707–721. Serrats, J., Schiltz, J.C., Garcı´a-Bueno, B., van Rooijen, N., Reyes, T.M., and Sawchenko, P.E. (2010). Dual roles for perivascular macrophages in immuneto-brain signaling. Neuron 65, 94–106. Sharma, K., Schmitt, S., Bergner, C.G., Tyanova, S., Kannaiyan, N., ManriqueHoyos, N., Kongi, K., Cantuti, L., Hanisch, U.K., Philips, M.A., et al. (2015). Cell

type- and brain region-resolved mouse brain proteome. Nat. Neurosci. 18, 1819–1831. Sousa, C., Golebiewska, A., Poovathingal, S.K., Kaoma, T., Pires-Afonso, Y., Martina, S., Coowar, D., Azuaje, F., Skupin, A., Balling, R., et al. (2018). Singlecell transcriptomics reveals distinct inflammation-induced microglia signatures. EMBO Rep. 19, e46171. Tay, T.L., Mai, D., Dautzenberg, J., Ferna´ndez-Klett, F., Lin, G., Sagar, Datta, M., Drougard, A., Stempfl, T., Ardura-Fabregat, A., et al. (2017). A new fate mapping system reveals context-dependent random or clonal expansion of microglia. Nat. Neurosci. 20, 793–803. Thion, M.S., Low, D., Silvin, A., Chen, J., Grisel, P., Schulte-Schrepping, J., Blecher, R., Ulas, T., Squarzoni, P., Hoeffel, G., et al. (2018). Microbiome Influences Prenatal and Adult Microglia in a Sex-Specific Manner. Cell 172, 500– 516.e516. Tu, P.H., Gurney, M.E., Julien, J.P., Lee, V.M., and Trojanowski, J.Q. (1997). Oxidative stress, mutant SOD1, and neurofilament pathology in transgenic mouse models of human motor neuron disease. Lab. Invest. 76, 441–456. Van Hove, H., Martens, L., Scheyltjens, I., De Vlaminck, K., Pombo Antunes, A.R., De Prijck, S., Vandamme, N., De Schepper, S., Van Isterdael, G., Scott, C.L., et al. (2019). A single-cell atlas of mouse brain macrophages reveals unique transcriptional identities shaped by ontogeny and tissue environment. Nat. Neurosci. 22, 1021–1035. Voet, S., Prinz, M., and van Loo, G. (2019). Microglia in Central Nervous System Inflammation and Multiple Sclerosis Pathology. Trends Mol. Med. 25, 112–123. Wang, Y., Szretter, K.J., Vermi, W., Gilfillan, S., Rossini, C., Cella, M., Barrow, A.D., Diamond, M.S., and Colonna, M. (2012). IL-34 is a tissue-restricted ligand of CSF1R required for the development of Langerhans cells and microglia. Nat. Immunol. 13, 753–760. €n, N., Riviere, Weinhard, L., Neniskyte, U., Vadisiute, A., di Bartolomei, G., Aygu L., Zonfrillo, F., Dymecki, S., and Gross, C. (2018). Sexual dimorphism of microglia and synapses during mouse postnatal development. Dev. Neurobiol. 78, 618–626. Wlodarczyk, A., Holtman, I.R., Krueger, M., Yogev, N., Bruttger, J., Khorooshi, R., Benmamar-Badel, A., de Boer-Bergsma, J.J., Martin, N.A., Karram, K., et al. (2017). A novel microglial subset plays a key role in myelinogenesis in developing brain. EMBO J. 36, 3292–3308. Zeisel, A., Mun˜oz-Manchado, A.B., Codeluppi, S., Lo¨nnerberg, P., La Manno, G., Jure´us, A., Marques, S., Munguba, H., He, L., Betsholtz, C., et al. (2015). Brain structure. Cell types in the mouse cortex and hippocampus revealed by single-cell RNA-seq. Science 347, 1138–1142. Zhan, L., Krabbe, G., Du, F., Jones, I., Reichert, M.C., Telpoukhovskaia, M., Kodama, L., Wang, C., Cho, S.H., Sayed, F., et al. (2019). Proximal recolonization by self-renewing microglia re-establishes microglial homeostasis in the adult mouse brain. PLoS Biol. 17, e3000134. Zhong, S., Zhang, S., Fan, X., Wu, Q., Yan, L., Dong, J., Zhang, H., Li, L., Sun, L., Pan, N., et al. (2018). A single-cell RNA-seq survey of the developmental landscape of the human prefrontal cortex. Nature 555, 524–528.

Cell Reports 30, February 4, 2020 1281