Microglial subtypes: diversity within the microglial community

Review Microglial subtypes: diversity within the microglial community Vassilis Stratoulias 1 , Jose Luis Venero 2,3 , Marie-Ève Tremblay 4,5 & Bertrand Joseph 1,* Abstract Microglia are brain-resident macrophages forming the first active immune barrier in the central nervous system. They fulfill multiple functions across development and adulthood and under disease conditions. Current understanding revolves around microglia acquiring distinct phenotypes upon exposure to extrinsic cues in their environment. However, emerging evidence suggests that microglia display differences in their functions that are not exclusively driven by their milieu, rather by the unique properties these cells possess. This microglial intrinsic heterogeneity has been largely overlooked, favoring the prevailing view that microglia are a single-cell type endowed with spectacular plasticity, allowing them to acquire multiple phenotypes and thereby fulfill their numerous functions in health and disease. Here, we review the evidence that microglia might form a community of cells in which each member (or “subtype”) displays intrinsic properties and performs unique functions. Distinctive features and functional implications of several microglial subtypes are considered, across contexts of health and disease. Finally, we suggest that microglial subtype categorization shall be based on function and we propose ways for studying them. Hence, we advocate that plasticity (reaction states) and diversity (subtypes) should both be considered when studying the multitasking microglia. Keywords disease; heterogeneity; homeostasis; microglia; subtypes DOI 10.15252/embj.2019101997 | Received 13 March 2019 | Revised 29 April 2019 | Accepted 3May 2019 | Published online 2August 2019 The EMBO Journal (2019)38:e101997 Introduction Microglia were introduced to the scientific literature a century ago (Rı ´o-Hortega, 1919a,b,c; Fig 1). During normal physiological conditions, microglial cells with a ramified morphology are regularly distributed throughout the central nervous system (CNS; Rı ´o-Hortega, 1919b). Upon pathology, microglia transform their morphology and function, leading to propose a cascade of “reaction” from ramified to hypertrophic and ameboid phenotypes that still orients research today (Flanary et al, 2007; Graeber, 2010; Fig 1). With the recent advances in genetic tools allowing for fate mapping (Ginhoux et al, 2010), microglia are now considered to be tissue-resident macrophages of the CNS that arise exclusively from the embryonic yolk sac (Alliot et al, 1999; Schulz et al,2012;Kierdorfet al, 2013; Perdiguero et al, 2015). Microglia colonize the murine CNS from embryonic day (E)9.5 (Tay et al, 2017c) and represent a self-maintaining and long-lived cell population that persists for months, if not the entire lifespan of the organism (Lawson et al, 1992; Ajami et al, 2007, 2011; Mildner et al, 2007; Askew et al,2017;Fu ¨ger et al, 2017; Re ´uet al, 2017; Tay et al, 2017b). Beyond microglia functioning as mediators of injury, inflammation, and neurodegeneration, several roles in the healthy brain have been identified at an exponential rate this past decade (Cartier et al, 2014; Tremblay et al, 2015; Fig 1). Microglia exhibit widely differing functions depending on the stage of life, CNS region, and context of health or disease. Differences in microglial number, morphology, and gene expression were also reported between sexes (Schwarz et al, 2012; Crain et al, 2013; Lenz et al, 2013; Pimentel-Coelho et al, 2013; Butovsky et al, 2015; Dorfman et al, 2017; Hanamsagar et al, 2017; Krasemann et al, 2017). Adequate microglial functions are crucial for plasticity and behavioral adaptation to the environment (Salter & Stevens, 2017; Tay et al, 2017a). Throughout life, microglia contribute to neurogenesis, neuronal circuit shaping, vascular formation and remodeling, and maintenance of homeostasis (Tay et al, 2017c). During aging and in diseases, these cells may become reactive or impaired in their surveillance and phagocytosis (Streit, 2002; Koellhoffer et al, 2017; Spittau, 2017). Microglial contribution to diseases is associated with compromised physiological roles (e.g., in synaptic maintenance and plasticity; Tay et al, 2017a) and processes that are adaptive in the healthy brain, yet leading to cell death and tissue damage in pathological settings (e.g., excitotoxicity, oxidative stress, and inflammation; Weil et al, 2008). Microglial reaction can be triggered by any kind of insults or disturbances to the CNS. Persisting microglial reaction, associated often with proliferation, is involved in pathological conditions ranging from neurodevelopmental disorders, traumatic injuries, infectious diseases, tumors, and psychiatric disorders, to neurodegenerative diseases. 1Toxicology Unit, Institute of Environmental Medicine, Karolinska Institutet, Stockholm, Sweden 2Departamento de Bioquímica y Biología Molecular, Facultad de Farmacia, Universidad de Sevilla, Sevilla, Spain 3Instituto de Biomedicina de Sevilla-Hospital Universitario Virgen del Rocío/CSIC/Universidad de Sevilla, Sevilla, Spain 4Department of Molecular Medicine, Université Laval, Quebec, QC, Canada 5Axe Neurosciences, Centre de Recherche du CHU de Québec-Université Laval, Quebec, QC, Canada *Corresponding author. Tel: +46 703057405; E-mail: bertrand[email protected] ª2019 The Authors. Published under the terms of the CC BY 4.0license The EMBO Journal 38:e101997 |2019 1of 18 Depending on the stage of the life, CNS region, and stressor or pathological insult at play, the microglial reaction process was shown to proceed differently and to result in sometimes contrasting outcomes (see Fig 2A for a classical schematic representation, depicting a ramified gray microglial cell surrounded by a palette of colorful microglia each representing a distinct reaction state). It is also now recognized that microglia display a wide range of reaction states, a tremendous shift from the M1/M2 classification still used a few years ago (Martinez & Gordon, 2014; Ransohoff, 2016). According to this view, the numerous functions of microglia would be fulfilled through their reaction toward multiple phenotypes, each associated with a distinct molecular signature (Crain et al, 2013; Hickman et al, 2013; Butovsky et al, 2014; Bennett et al, 2016; Grabert et al, 2016; Flowers et al, 2017; Galatro et al, 2017; KerenShaul et al, 2017; Krasemann et al, 2017; Hammond et al, 2018; Masuda et al, 2019). However, several pieces of evidence also indicate that different pools of microglia might each display distinct intrinsic properties that would be acquired during their maturation or function within the CNS. These subtypes would co-exist at steady state and undergo further modulation or phenotypic transformation in response to stimuli (Fig 2B). Indeed, beyond the view that microglia are a unique cell type in the CNS that adopts different phenotypes in response to different stimuli, we propose in this review article that microglia might constitute a community of cells in which different members display distinct properties, perform distinct physiological functions, and respond differently to stimuli (Fig 2C). We review the distinctive features of several putative microglial “subtypes”, at structural, ultrastructural, and expression levels, as well as their functional implications across contexts of health and disease. Furthermore, we propose to categorize microglial subtypes based on functions, rather than molecular signatures and markers. Finally, we suggest that microglial subtype candidates should be validated using a methodological workflow that we recommend. Microglia: a community fulfilling the vast microglial functions What defines a cell subtype is subject of intense debate, and it is discussed in Box 1. Accumulating evidence indicates that microglia 2019 First systematic attempts to identify microglial subtypes 2018 Alternative microglia ontogeny for Hoxb8+ microglia 2017 Description of neurodegenerative diseases associated microglia (DAM) 2017 Subpopulation of myelinogenic CD11c+ microglia 2016 Description of dark microglia 2015 In zebrafish: distinct origins for embryonic and adult microglia 2014 Subpopulation of neurogenic microglia in SVZ/RMS 2014 Mixed responses of microglia to neuro-transmitters/hormones 2012 Mixed responses of microglia to LPS 1993 Subpopulation of KS+ microglia 2002 Subpopulation of TREM2– microglia 2010 Subpopulation of Hoxb8+ microglia 1919 Description of “satellite microglia” 1919 1920s 1930s 1940s 1950s 1960s 1970s 1980s 1990s 2000s 2010s 2013–19 Genome-wide definition of distinctive microglial gene signatures in steady and reactive states 1919 Microglia identification; introduction of amyboid vs ramified cascade of microglial “activation” 2010 Yolk sac origin for microglia 2005 Microglia perform vital functions in steady state 2020s 1990s 2000s 2020s 1993 2002 2005 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2011 Subpopulation of exosome uptake-competent microglia © EMBO Figure 1. Historical overview of microglial subtype identification. Although microglial subtypes have originally been proposed by Rio-Hortega in the first report of microglia, it was only recently that this idea was revisited. 2of 18 The EMBO Journal 38:e101997 |2019 ª2019 The Authors The EMBO Journal Vassilis Stratoulias et al are not the naı ¨ve cell type that invariably responds identically to any possible type of stimuli by assuming a predetermined phenotype. In fact, from a historical perspective, the notion of microglial subtypes had already been proposed in 1919 by Rio-Hortega in his original description of microglia (Rı ´o-Hortega, 1919b; Fig 1). He noticed that some microglia that he named “satellite” microglia were found in close proximity to neuronal cell bodies. A century later, we propose that the satellite microglia, which are discussed below, might represent one of the playing cards in the deck of microglial subtypes (Fig 3). It is important to acknowledge that others, avant-garde scientists, have paved the way for the concept of microglial diversity (McCluskey & Lampson, 2000; Olah et al, 2011; Hanisch, 2013; Gertig & Hanisch, 2014). Microglial regional heterogeneity at steady state Although microglia are ubiquitously scattered throughout the CNS, their distribution varies across regions, also between the white matter and gray matter (Lawson et al, 1990). Microglial morphology differs with the presence of neuronal cell bodies, dendrites and axons, myelinated axons, and blood vessels. Furthermore, microglia exhibit regional differences in self-renewal and turnover rates under normal physiological conditions and upon stimuli, such as lipopolysaccharide (LPS) challenge (Lawson et al, 1992; Ajami et al, 2007, 2011; Mildner et al, 2007; Askew et al, 2017; Fu ¨ger et al, 2017; Re ´uet al, 2017; Tay et al, 2017b; Furube et al, 2018). The regional microenvironment has been shown to tightly determine microglial identity at the transcriptional level, in both mouse and human (Gosselin et al, 2014, 2017). Direct evidence for microglial regional variability notably comes from studies in which microglia were isolated from wild-type, unchallenged adult mice, according to brain area, and their transcriptome was determined based on panels of pre-selected microglial markers. In one study, the expression of CD11b, CD40, CD45, CD80, CD86, F4/80, TREM2b, CX3CR1, and CCR9 was compared among microglia isolated from different CNS A B C P a t h o l o g i c a l c o n d i t i o n s D e v e l o p m e n t A g i n g P a t h o l o g i c a l c o n d i t i o n s P a t h o l o g i c a l c o n d i t i o n s D e v e l o p m e n t A g i n g P a t h o l o g i c a l c o n d i t i o n s Young Newborn Fetus Adult Aged Disease 1 Disease 2 Disease 3, … Injury 1,… Disorder 2,… Disorder 1 Young Newborn Fetus Adult Aged Disease 1 Disease 2 Disease 3, … Injury 1,… Disorder 2,… Disorder 1 Steady state Environmental cue Non-responsive Expand Change © EMBO Figure 2. Microglial reaction states. (A) Currently, microglia are considered a homogenous cellular population (core of the circle in gray) that is extremely plastic. Depending on the brain homeostasis status at a given developmental stage or resulting from pathology, microglia respond invariably to assume a wide range of phenotypes as described in the literature. (B) In the updated version proposed here, microglia constitute a heterogeneous cell population having intrinsic properties and functional specializations. (C) Upon an environmental cue, each microglial subtype may respond or not to the stimulus, by expanding and/or changing its morphology and gene expression to assume a specific phenotype. ª2019 The Authors The EMBO Journal 38:e101997 |2019 3of 18 Vassilis Stratoulias et al The EMBO Journal regions of young adult mice (de Haas et al, 2008). Although all of these markers were expressed across the CNS, their protein expression varied significantly between areas. In a similar study performed in adult rats, the expression levels of known microglial markers also showed region-specific profiles (Doorn et al, 2015). Similar studies performed in mice that compared microglia isolated from different brain areas additionally showed regional heterogeneity in expression pattern throughout the lifespan (Butovsky et al, 2014; Grabert et al, 2016; De Biase et al, 2017; Masuda et al, 2019). Additionally, in an unbiased single-cell RNA sequencing (RNAseq) study, in which cerebral tissue and hippocampal tissue from unchallenged young adult mice were analyzed, 47 molecularly distinct cell subtypes were identified, including two belonging to the microglia (Zeisel et al, 2015). These findings raise the intriguing possibility that regional differences in terms of neuronal survival, activity, growth factor release, metabolism, as well as synaptic plasticity, myelination, vascular remodeling, blood–brain barrier properties, may require distinct microglial functions, thus driving the differentiation of distinct microglial subtypes during development or function within the CNS. These microglial subtypes could be a major contributing factor to the microglial regional heterogeneity. Recently, cerebellar microglia were shown to display a unique clearance ability, defined by their expression of numerous genes supporting the engulfment and catabolism of cells or cellular debris (Ayata et al, 2018). This cerebellar microglial “type” is reminiscent of developing microglia and disease-associated microglia (DAM) that will be discussed below. By contrast, microglia from the striatum display a homeostatic surveillance phenotype. This microglial differentiation in response to regional differences in the environment was shown to be driven by epigenetic mechanisms (Ayata et al, 2018). In particular, the suppression of clearance genes in striatal microglia is mediated by PRC2, which catalyzes the repressive chromatin modification histone H3 lysine 27 trimethylation (H3K27me3). The ablation of PRC2 in microglia also results in the emergence of clearance microglia even in the absence of dying neurons, among both the striatum and cerebral cortex. These aberrant clearance microglia induce impaired motor responses, decreased learning and memory, together with the development of anxiety and seizures in mice (Ayata et al, 2018). A recent study that characterized the diversity of CNS-associated macrophages (CAM) also identified three different subsets of CAM that expressed high levels of Mrc1, Ms4at, Pf4, Stab1, Cbr2, CD163, and Fcrls, and were associated with different CNS compartments: the leptomeninges, choroid plexus, and perivascular space (Jorda ˜oet al, 2019). Consequently, some of the regional microglial diversity described using these markers could also be partly accounted for by CAM diversity. Microglial subtypes as defined by differential gene expressions Differential gene expression is an established approach for defining distinct subpopulations of a cell type, for instance the different neuronal subtypes (e.g., GABAergic and glutamatergic) observed in the healthy brain. In various contexts, neighboring microglia were shown to display differences in gene expression at steady state. These observed differences between microglia could arise from local cues, including interactions with different subtypes of neurons (e.g., inhibitory and excitatory) and glial cells (astrocytes, oligodendrocytes, and progenitors), or slight differences in signaling thresholds. Box 1: How to define a “cell (sub)type” The answer to “how to define a cell subtype?”is probably to be found in the answer to a closely related question, “how to define a cell type?”Traditionally, a cell type is defined based on its host tissue, morphology, lineage, function, and molecular composition. However, the definition of this term remains subject to intense debate (Clevers et al,2017). The advancement of unbiased technologies for single-cell transcriptome profiling, such as high throughput single-cell RNAseq and mass cytometry (or improved/related methods), has revealed remarkable heterogeneity among cells which were traditionally considered tobe homogeneous. However, whereas this degree of transcriptome and proteome heterogeneity is sufficient for defining cell subtypes, or even cellular states, is also a topic of intense debate (Trapnell, 2015; Okawa et al,2018). While single-cell RNAseq and mass cytometry allow to define molecularly distinct cell subpopulations, these approaches require to be complemented by the identification of the unique functions associated with these cell populations, in order to define those as cell (sub)types. Worth a notice, it is of importance not to confound cell subtypes with cellular states of reaction. The latter is referring to the different phenotypes and associated functions a cell type may acquire in response to various stimuli. A cell subtype should be defined by shared properties/characteristics within other cells within the cell type. Their unique intrinsic features and selective physiological functions should also be independent from their microenvironment. These two concepts are not mutually exclusive, as a cell subtype in response to a stimulus could react and acquire a new phenotype, i.e., reaction states, thus adding another level of complexity. Microglial subtypes must be defined in steady-state and unchallenged conditions by their intrinsic propertie(s) which translate into unique physiological function(s). Typically, the existing literature is the foundation of a research plan, which by definition is biased in respect to studies aimed at identifying a new cell type or subtype. This includes any work with markers, most importantly staining, sorting, and isolation of cells. Reverse genetic approaches can provide a more reliable tool for such studies, but still they have inherit technical limitations such as cell gating in flow cytometry and antibody unspecificity (Luo et al,2013). On the other hand, unbiased technologies such as single-cell RNAseq, mass cytometry, and electron microscopy are useful tools, but still we should be aware of their limitation in terms of providing a static view of cellular dynamics. They however become useful when combined with twophoton in vivo imaging to provide insights into dynamics. Serendipitous identification is also an approach, but it is sporadic and by definition nonsystematic. All of the above methodologies can contribute to the identification of new microglial subtypes. Considering the various putative subtypes that we have discussed in this review, a need for classifying microglial subtypes is evident. Deciphering whether their variations are instructed by the microenvironment or whether they result from intrinsic properties is of prime importance, using the following methodological workflow: Fate-mapping strategies allowing to visualize selectively different microglial subsets, for instance using non-invasive chronic two-photon in vivo imaging—could be performed longitudinally across development, adulthood, and aging, under steady-state as well as disease conditions—to determine the identity of putative microglial subtypes as microglial subsets or phenotypes. Microglia could be considered subtypes if their defining properties remain when these cells are examined longitudinally, under steady-state or disease conditions. They would however be considered phenotypes if instead they can transform one into another, notably in response to stimuli. The molecular determinants and physiological roles of the distinct subsets could then be studied using a combination of gene and protein expression analyses, as well as morphology, ultrastructure, and dynamic investigations. 4of 18 The EMBO Journal 38:e101997 |2019 ª2019 The Authors The EMBO Journal Vassilis Stratoulias et al “Satellite microglia” Function Interact with the axon initial segment (AIS) of neurons in the healthy brain. Loss of interaction upon injury. Characteristics Localized to the axonal side of the neuron’s cell body exhibit a single process overlapping with the AIS. Defined as microglia IBA1 CD11b CX3CR1 Reference Del Río Hortega, 1919 Baalman et al, 2015 Wogram et al, 2016 Hippocampus Cortex Microglia Neuron KSPG-microglia Function Typical appearance upon different insults, i.e. around motorneurons in ALS; away from brain trauma Characteristics Keratan sulfate positive cells Defined as microglia IBA1 CR3 CD11b Reference Bertolotto et al, 1998, 1995 Hirano et al, 2013 Shinjo et al, 2014 HippocampusOlfactory bulb Microglia Brainstem Microglia supporting neurogenesis Function Essential for neuroblast survival and migration in SVZ/RMS Characteristics IBA1–, isolectin B4–, CD68-negative, P2RY12low, pSTAT6-positive cells, IL4 and IL10-producing cells, less ramified than microglia in neighbouring brain areas Defined as microglia CX3CR1-EGFP Reference Shigemoto et al, 2014 Ribeiro Xavier et al, 2015 Xavier et al, 2015 Olfactory bulb Subventricular zone Rostral migratory stream Microglia Neuroblasts Hox8b-microglia Function Absence of Hoxb8 function in microglia impacts on the corticostriatal neuronal circuit, and leads to impaired grooming, anxiety and social behaviors Characteristics YFP-Hoxb8-expressing cells Defined as microglia IBA1 Cd11b Reference Chen et al, 2000 Nagrajan et al, 2017 De et al, 2018 Cortex Olfactory bulb Microglia Cortico-striatal neuron Cd11c-microglia Function Promote myelination and neurogenesis in the neonatal brain Characteristics CD11c-positive, Igf1-producing cells Defined as microglia IBA1 CD11b CD45low CX3CR1 CCR2null Reference Wlodarczyk et al, 2017 Corpus callosum Cerebellum Microglia Myelin sheath Neuron Oligodendrocyte “Dark microglia” Function Interact with blood vessels and synapses Characteristics Appear as dark by Electron Microscopy Defined as microglia IBA1low CX3CR1-GFPlow CD11b TREM2 4D4 Reference Bisht et al, 2016 Hui et al, 2018 Hippocampus Cortex Amygdala Hypothalamus Dark microglia Figure 3. Putative microglial subtypes with unique specializations. Emerging data provide support to the existence of putative microglial subtypes endowed with unique genomic, spatial, morphological, and functional specializations. We anticipate that analyzing these subtypes thoroughly, with the methodological workflow proposed in Box 1, and using a similar methodology for newly discovered ones, will result in the identification of a number of different microglial subtypes with unique functional characteristics that could be targeted for disease prevention or treatment. ª2019 The Authors The EMBO Journal 38:e101997 |2019 5of 18 Vassilis Stratoulias et al The EMBO Journal Similarly, differences in peripheral macrophage activation by LPS and viruses have been described, where only a subset of the population concomitantly displays a response (Ravasi et al, 2002). In addition, microglia may directly communicate with each other, which suggests that the recruitment of a specific microglial cell might lead to an inhibition of the neighboring microglia. Microglia were initially defined as occupying non-overlapping territories in the healthy brain, but this view is now changing, with improved staining methods showing direct contacts between processes and sometimes cell bodies from neighbor microglial cells (for example, see Milior et al, 2016). Furthermore, the possibility that differential marker expression among adjacent microglia results from differences in microglial Pathological conditions Canonical microglia TOOLBOX Non-canonical microglia i.e. microglia subtype candidate Identified by • differential gene/protein(s) expression • structural/ultrastructural properties • unique functions • others in the wild-type and unchallenged CNS Canonical microglia Non-canonical microglia Addition of a new microglia subtype Experiments to be performed prior to classification as subtypes Additional experiments of specific interest Gene expression profiles Transcriptomic analysis Isolation Time Time Time Transcriptomic comparisons of canonical versus non-canonical microglia Ontogeny Lineage tracing analysis Spatial & temporal distribution Confocal, 3D imaging Functional studies In vitro studies In vivo studies Population dynamics Migration, apoptosis, proliferation Control mouse Diseased mouse Induction of linage tracer expression × loxP loxP loxP Cre Reporter Reporter Rosa26 Rosa26 Subtype-specific gene Cre expression STOP Cx3cr1 II III +Tamoxifen IV V VIExon Exon loxP loxP loxP VVI VI Cre ERT2 × Phenotypic analysis and behavioral studies Yolk sac ? ? ? ? ? ? © EMBO Figure 4.Toolbox. 6of 18 The EMBO Journal 38:e101997 |2019 ª2019 The Authors The EMBO Journal Vassilis Stratoulias et al exposure to previous challenges also has to be considered. For instance, it has been shown using non-invasive two-photon in vivo imaging that neighbor microglia respond differently to laser injury in the intact, unchallenged brain, leading to their processes converging or not toward the site of injury (Nimmerjahn et al, 2005; Paris et al, 2018). In addition, microglial cell bodies were recently shown to migrate in the cerebral cortex (Eyo et al, 2018) and cerebellum (Stowell et al, 2018) of healthy adult mice, which paints another layer of complexity. However, the existence of microglial subtypes, each endowed with intrinsic differences in gene expression, cannot be excluded and we argue that the topic deserves further investigation. Putative microglial subtypes are discussed below: Keratan sulfate proteoglycan (KSPG)-microglia A quarter of century ago, microglia were shown in the unchallenged adult rat brain to exhibit constitutive heterogeneity in their expression of KSPG (Bertolotto et al, 1993), visualized in situ using the 5D4 monoclonal antibody (Fig 3). KSPG is located in the extracellular matrix and on the cell surface. They are suggested to contribute to the control of cellular adhesion and axonal growth. In particular, 5D4-KSPG is expressed by a subpopulation of ramified microglia, contrary to ameboid microglia and peripheral macrophages (Bertolotto et al, 1993, 1998). Of note, 5D4-KSPG expression does not coincide with the expression of GFAP, NG2, or MAP2, which relate to other CNS cells. The expression of 5D4-KSPG in microglia differs significantly between strains of inbred rats (Jander & Stoll, 1996b). In mammals, a subpopulation of 5D4-KSPG-expressing microglia was also reported in the spinal cord and retina (Bertolotto et al, 1993, 1998; Jander & Stoll, 1996a; Jones & Tuszynski, 2002; Zhang et al, 2005; Foyez et al, 2015). The 5D4-KSPG-microglia exhibit a preferential regional distribution in the CNS. Indeed, whereas these cells are found in large numbers among the hippocampus, brainstem, and olfactory bulb (OB), only few of them are detected in the cerebellum and cerebral cortex (Bertolotto et al, 1993, 1998). This putative microglial subset is also observed in the neonatal rat brain (Bertolotto et al, 1998). It is of importance to mention that 5D4KSPG-microglia were shown to co-exist with 5D4-KSPG-negative microglia in the same CNS regions (Jones & Tuszynski, 2002). Although these studies argue for the presence of two different subtypes, based on KSPG-reactivity, a systematic approach is required to confirm this possibility (Fig 4). Hox8b-microglia These microglial cells have a molecular signature that differentiates them from the canonical population, together with a unique spatial and temporal distribution (see Box 2 for distinct ontogeny of Hoxb8microglia). Mice carrying the driver Hoxb8-Cre and the reporter ROSA26-YFP alleles were crossed to trace YFP-Hoxb8 expression. In the adult brain, the only cells showing YFP signal appeared to be microglia. YFP-positive microglia were found throughout the brain, especially in the cerebral cortex and OB (Chen et al, 2010; De et al, 2018; Fig 3). YFP-positive microglia, which represent 25–40% of the total microglial population in the adult brain, were also shown to co-exist with YFP-negative microglia (Chen et al, 2010; De et al, 2018; Nagarajan et al, 2018). Transcriptomic analyses comparing Hoxb8-positive and Hoxb8-negative microglia revealed that they are very similar at steady state, with only 21 genes differing significantly in expression between the two populations (De et al, 2018). Hoxb8-microglia express microglial signature genes, such as Tmem119,Sall1,Sall3,Gpr56, and Ms4a7, and genes associated with hematopoietic ontogeny including Clel12a,Klra2, and Lilra5 at similar levels compared with non-Hoxb8 canonical microglia (Bennett et al, 2018; De et al, 2018). Of note, neither of the two putative microglial subtypes was found to expresses Hoxb8 in the adult brain; instead, the lineage tracer approach revealed that Hoxb8 is expressed by microglial progenitors prior to CNS infiltration (De et al, 2018). Selective inactivation of Hoxb8 in the hematopoietic system was also sufficient to induce pathological grooming behavior, as observed in constitutive Hoxb8 mutant mice (Chen et al, 2010; Nagarajan et al, 2018). The strategy for gene deletion included the use of Tie2 Cre mice that affect all hematopoietic cells and endothelial cells (Chen et al, 2010). More cell-specific deletion of Hoxb8 within microglial cells is a prerequisite to determine their selective involvement in pathological grooming behavior. CD11c-microglia Recently, evidence for an additional microglial subtype expressing the integrin CD11c was uncovered in neonatal mouse brain (Fig 3). CD11c-microglia expand during postnatal development to represent approximately one-fifth of the total microglial population at postnatal day(P)3-P5, and then drop to represent less than 3% of the population in juvenile and adult mice (Wlodarczyk et al, 2017). Whether this decrease in numbers is due to increased cell death, migration or even trans-differentiation of this putative microglial subset is yet to be determined. These microglia distribute unevenly across the developing brain, being found predominantly in primary myelinating regions, mainly the corpus callosum and cerebellar white matter. Gene ontology enrichment analysis revealed that CD11cmicroglia express genes associated with neurogenic and myelinogenic processes in the neonatal brain. They are also a major source of insulin-like growth factor 1 (IGF1), while selective depletion of IGF1 in this microglial subtype leads to impaired developmental myelination (Wlodarczyk et al, 2017). Thus, CD11c-microglia in the neonatal mouse brain appear to play essential roles in neurogenesis and myelinogenesis during development. TREM2-microglia There is additional evidence that differential microglial gene expression patterns define several microglial populations in the healthy brain. For example, not all microglia express the cell surface receptor TREM2, which is known to impact on their survival, proliferation, clustering around amyloid-beta (Ab) plaques in Alzheimer’s disease (AD) pathology, phagocytosis, and metabolism (Yeh et al, 2017). Furthermore, loss-of-function variants in the TREM2 protein increase the risk of developing late-onset AD among other forms of dementia (Colonna & Wang, 2016). However, despite the apparent essential function of TREM2 in microglia, its expression is far from being ubiquitous and homogenous in those cells throughout the brain. In the mouse brain, microglial expression of TREM2 also varies between and within individual regions (Schmid et al, 2002). The numbers of TREM2-expressing cells are highest in the cingulate cortex and lateral entorhinal cortex, and much lower in the hypothalamus and habenula, while some regions, such as the circumventricular organs, completely lack TREM2 expression. Worth notice, even in brain regions abundant in TREM2-expressing microglia, TREM2positive and TREM2-negative cells were found in the immediate ª2019 The Authors The EMBO Journal 38:e101997 |2019 7of 18 Vassilis Stratoulias et al The EMBO Journal vicinity to each other (Schmid et al, 2002). Regional differences in TREM2 gene expression, associated with microglial markers, are also observed in the human brain (Forabosco et al, 2013). Microarray data generated from 101 healthy control individuals reveal significant regional differences in TREM2 gene expression between the white matter and cerebellum (Forabosco et al, 2013). This regional heterogeneity in TREM2 expression could point toward a specific subtype. Microglia supporting neurogenesis The use of CX3CR1-EGFP reporter mice, in which one of the loci of fractalkine receptor Cx3cr1 is replaced by the gene encoding EGFP, revealed heterogeneity of the microglial cell population within the neurogenic subventricular zone (SVZ), and the adjacent rostral migratory stream (RMS) that terminates into the OB (Ribeiro Xavier et al, 2015; Xavier et al, 2015; Fig 3). First, it was noticed that microglia located along the SVZ-RMS-OB axis are significantly less ramified than microglia from adjacent areas in mice and rats (Shigemoto-Mogami et al, 2014; Ribeiro Xavier et al, 2015; Xavier et al, 2015). CX3CR1-EGFP expressing microglia were TREM2-negative, and about half of them also IBA1-negative in the SVZ and RMS of wild-type adult mice. By contrast, CX3CR1-EGFP expressing microglia expressed TREM2 in the OB of wild-type adult mice. Further adding to this diversity, a subset of the latter population additionally expressed CD68 (~35%) and/or isolectin B 4 (~15%) in the OB (Ribeiro Xavier et al, 2015). Significantly increased CD68 expression in microglia was also reported in the OB of adult wildtype rats (Doorn et al, 2015), while a regional distribution of CD68positive microglia was reported in the brain of human midterm fetuses (gestational ages of 15–25 weeks; Cho et al, 2013). As observed in the SVZ and RMS, IBA1-negative and IBA1-positive microglia were also observed in close vicinity to another within the OB. Of note, the antigenic heterogeneity of CX3CR1-EGFP expressing microglia was reported in later development among the SVZ of newborn (P1) and early postnatal (P7) mice, suggesting persistence of these microglial cell populations beyond ontogeny (Xavier et al, Box 2(with associated illustration): Revisiting the microglial origin(s) An important question arising from the existence of microglial subtypes relates to their possible origin(s). Do microglial subtypes possess intrinsic differences prior to populating the CNS, or do they acquire their unique properties once they have assumed their regional distribution within the CNS parenchyma? Current literature states convincingly that microglia derive from the first wave of hematopoiesis from the embryonic yolk sac in mouse (Ginhoux et al,2010; Hoeffel et al,2015; Perdiguero et al,2015; Sheng et al,2015; Mass et al,2016), where they follow a stepwise maturation program (Mass et al,2016; Matcovitch-Natan et al,2016), before populating the embryonic brain at E9.5(Tay et al,2017c). Based on the above literature, microglial subtypes should differentiate once they have assumed their regional distribution inside the CNS parenchyma (a). This hypothesis could explain microglial differences resulting from regional differences in microenvironments or from differences in local cues among the microenvironment such as microglial interactions with different neuronal subtypes (inhibitory, excitatory) and glial cells (astrocytes, oligodendrocytes and their progenitors), or slight differences in signaling thresholds, leading to the observed differences in adjacent microglia. The alternative hypothesis which is based on microglial cells exhibiting intrinsic differences prior to infiltrating the CNS cannot be excluded at this early stage of investigation, and should be tested (b and c). In support of the later hypothesis, Capecchi et al reported that Hoxb8-microglia-progenitors already exist in the yolk sac at E8.5(De et al,2018). Subsequently, these cells transit through the aorta-gonadmesonephros and fetal liver, where they expand in number, prior to their entry into the brain at E12.5(De et al,2018) (c). On the same lines, microglial cells found in CSF1R / (Ginhoux et al,2010; Erblich et al,2011) and in IL2-Tgfb1;Tgfb1 / (Keren-Shaul et al,2017) transgenic mice are expected to exhibit intrinsic differences prior to infiltrating the brain parenchyma. Recently, it has been reported that at E14.5two microglial subpopulations exist, based on Ms4a7expression (Hammond et al,2018). It would be of great interest to investigate the ontogeny of these two subpopulations. In zebrafish, two waves of microglial infiltration have been reported (Xu et al, 2015; Ferrero et al,2018). Microglia of a yolk-sac-equivalent structure origin initially populate the embryonic brain. Subsequently, the microglial population is replenished by adult microglia that derive from a distinct tissue later during zebrafish development (d). This microglial diversity could result from species-specific differences between zebrafish and mouse. Nevertheless, these studies indicate that evolutionarily multiple microglial origins and maturation programs are a possibility. Recently, it has been reported that at steady state a wave of monocytes infiltrate the mouse brain parenchyma at early postnatal stages; however, these cells were rapidly depleted and did not contribute to the later microglial population (Askew et al,2017). Development Embr y o A B C D Adult © EMBO 8of 18 The EMBO Journal 38:e101997 |2019 ª2019 The Authors The EMBO Journal Vassilis Stratoulias et al 2015). In addition, at those developmental ages, CD68 expression was detected in a microglial subset only. It would be of particular interest to determine whether the CD68-expressing microglia found in the OB of adult mice originate from the ones detected in the SVZ of newborn mice and whether the microglial heterogeneity described above translates into one or more microglial subtypes during adulthood and aging. A recent study has taken advantage of single-cell RNAseq analysis to uncover regional microglial heterogeneity across different brain regions (Li et al, 2019). Contrary to expected, the authors found a remarkable homogeneity of adult homeostatic microglia (enriched in homeostatic genes such as TMEM119 or P2ry12) regardless of the brain tissue of origin. However, a microglial population expressing Clec7a, one of the most upregulated genes in DAM (Keren-Shaul et al, 2017; Krasemann et al, 2017), was shown by immunohistochemistry to be restricted to the subgranular zone of the hippocampal dentate gyrus, another neurogenic niche, in addition to the SVZ and RMS. This restrictive pattern of localization suggests the existence of a microglial subtype that exerts a key role in adult neurogenesis. These findings illustrate the usefulness of combining different methodological strategies to uncover novel microglial populations among specific CNS regions. These selected examples of differential gene expression among small populations of microglia argue for the existence of distinct microglial subtypes. Of importance, in the majority of these cases, the putative microglial subtypes co-existed with canonical microglia within the same microenvironment. While extrinsic factors such as local cues in the microenvironment, as discussed above, could determine this variability, microglia also appear to have intrinsic differences that warrant further investigation. Single-cell RNAseq data Recently, there has been an unprecedented influx of data from single-cell RNAseq studies in support of a significant microglial heterogeneity. Although at this point these pieces of evidence are descriptive and we suggest that they have to be coupled with functional studies to categorize the identified clusters as different microglial subtypes, it is striking to report such a plethora of different potential microglial subtypes. In particular, Stevens and colleagues performed deep single microglial cell RNAseq at different developmental stages and uncovered eight transcriptionally distinct microglial clusters co-existing in the naı ¨ve mouse brain (Hammond et al, 2018). Among them, “cluster 2” is characterized by a high expression of Ms4a family members, some of which are involved in immune cell functions (Eon Kuek et al, 2016), and partially overlaps with brain border macrophage markers (Hammond et al, 2018). The number of cells belonging to “cluster 2” decreases drastically during postnatal development. “Cluster 3” is almost exclusively found in the embryonic and early postnatal brain. It is characterized by its unique expression of Fabp5, while “cluster 6” highly expresses genes that include Cd74,Ccl24, and Arg1 and is enriched in female samples (Hammond et al, 2018). “Cluster 4” [named axonal tract-associated microglia (ATM) by the authors] has a very specific spatiotemporal expression. Microglia belonging to this cluster highly express Spp1,Gpnmb, Igf1,CD68, and Lgals3, and they have an ameboid morphology. Furthermore, their number is significantly enriched in the neonatal brain and they are preferentially localized in the white matter, including the corpus callosum, and cerebellum (Hammond et al, 2018). Proliferative-region-associated microglia (PAM) Similarly, Barres and colleagues performed deep single microglial cell RNAseq at different developmental stages including late embryonic, early postnatal (P7) and adult ones across different brain regions (Li et al, 2019). Three well-defined microglial clusters irrespective of cell cycle states were identified during postnatal development (Li et al, 2019) and initially named P7-C0, P7-C1, and P7C2. The P7-C0 and P7-C1 microglial clusters both expressed homeostatic genes although at lower levels for P7-C1. This specific cluster expressed many genes recently identified in DAM including Igf1,Spp1,Gpnmb,CD11c (also known as Itgax), and Clec7a (Keren-Shaul et al, 2017; Krasemann et al, 2017). Further analysis identified these cells to be predominantly located in the corpus callosum and cerebellar white matter, where they intermingled with Mbp + oligodendrocytes; hence, they were named the proliferative-region-associated microglia (PAM). These PAM are highly reminiscent of the CD11c-microglia described by Wlodarczyk et al (2017) (Wlodarczyk et al, 2017) and the ATM cluster identified by Hammond et al (2018). Indeed, PAM populate the white matter at P4, peak at P7, and almost disappear by P14. At the morphological level, PAM are ameboid with thicker primary branches and larger cell bodies, as compared with the typical microglia. They are also highly phagocytic, and contrary to the DAM, they do not depend on either triggering receptor expressed on myeloid cells 2 (TREM2) or ApoE (Keren-Shaul et al, 2017; Krasemann et al, 2017). Although it is clear that postnatal PAM may represent a specialized microglial subtype involved in the elimination of oligodendrocytes during myelination, the exact timing and appearance of this microglial population in the white matter during periods of myelination indicates that the environmental needs may be behind the polarization of homeostatic microglia toward PAM. In a third study, Prinz and colleagues (Masuda et al, 2019) analyzed single microglial cells derived from brain areas that were previously shown to exhibit regional transcriptional differences (Grabert et al, 2016), namely cortex, cerebellum, and hippocampus, and additionally analyzed microglia from corpus callosum. Microglia from juvenile (3 weeks old) and adult (16 weeks old) animals were grouped into at least four distinct clusters which showed a variable distribution between brain regions and developmental stage (Masuda et al, 2019). Importantly, this study also identified multiple different microglial clusters in human tissue, from cortex that bears no signs of CNS pathology (Masuda et al, 2019), therefore indicating that microglial heterogeneity is also relevant to human (see also later). Microglial subtypes as defined by differential structural/ ultrastructural properties Returning to the satellite microglia described by Rı ´o-Hortega (1919b), this subpopulation is currently identified based on its unique feature of having its soma being associated with neuronal cell bodies (Baalman et al, 2015; Wogram et al, 2016; Fig 3). Half of the satellite microglia extend a single process that overlaps with the portion of the axon where potentials are initiated (Baalman et al, 2015). They have been identified both during development and at adulthood in mice, while they have a preferential association with excitatory neurons (Baalman et al, 2015). In addition, this subpopulation was reported in the cerebral cortex of adult rats and adult non-human primates (Rı ´o-Hortega, 1919b; Baalman et al, 2015), hence indicating conservation across species. Whether ª2019 The Authors The EMBO Journal 38:e101997 |2019 9of 18 Vassilis Stratoulias et al The EMBO Journal Habib N, Avraham-Davidi I, Basu A, Burks T, Shekhar K, Hofree M, Choudhury SR, Aguet F, Gelfand E, Ardlie K et al (2017) Massively parallel singlenucleus RNA-seq with DroNc-seq. Nat Methods 14:955 –958 Haimon Z, Volaski A, Orthgiess J, Boura-Halfon S, Varol D, Shemer A, Yona S, Zuckerman B, David E, Chappell-Maor L et al (2018) Re-evaluating microglia expression profiles using RiboTag and cell isolation strategies. Nat Immunol 19:636 –644 Hammond TR, Dufort C, Dissing-Olesen L, Giera S, Young A, Wysoker A, Walker AJ, Gergits F, Segel M, Nemesh J et al (2018) Single-cell RNA sequencing of microglia throughout the mouse lifespan and in the injured brain reveals complex cell-state changes. Immunity 50:253 –271 Han J, Harris RA, Zhang XM (2017) An updated assessment of microglia depletion: current concepts and future directions. Mol Brain 10:25 Hanamsagar R, Alter MD, Block CS, Sullivan H, Bolton JL, Bilbo SD (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 Hanisch UK (2013) Functional diversity of microglia - how heterogeneous are they to begin with? Front Cell Neurosci 7:65 Härtlova A, Erttmann SF, Raffi FA, Schmalz AM, Resch U, Anugula S, Lienenklaus S, Nilsson LM, Kröger A, Nilsson JA et al (2015) DNA damage primes the type I interferon system via the cytosolic DNA sensor STING to promote anti-microbial innate immunity. Immunity 42:332 –343 Hickman SE, Kingery ND, Ohsumi TK, Borowsky ML, Wang LC, Means TK, El Khoury J (2013) The microglial sensome revealed by direct RNA sequencing. Nat Neurosci 16:1896 –1905 Hirano K, Ohgomori T, Kobayashi K, Tanaka F, Matsumoto T, Natori T, Matsuyama Y, Uchimura K, Sakamoto K, Takeuchi H et al (2013) Ablation of keratan sulfate accelerates early phase pathogenesis of ALS. PLoS One 8: e66969 Hoeffel G, Chen J, Lavin Y, Low D, Almeida FF, See P, Beaudin AE, Lum J, Low I, Forsberg EC et al (2015) C-Myb(+) erythro-myeloid progenitor-derived fetal monocytes give rise to adult tissue-resident macrophages. Immunity 42:665 –678 Hoogland IC, Houbolt C, van Westerloo DJ, van Gool WA, van de Beek D (2015) Systemic inflammation and microglial activation: systematic review of animal experiments. J Neuroinflammation 12:114 Huang Y, Xu Z, Xiong S, Qin G, Sun F, Yang J, Yuan TF, Zhao L, Wang K, Liang YX et al (2018a) Dual extra-retinal origins of microglia in the model of retinal microglia repopulation. Cell Discov 4:9 Huang Y, Xu Z, Xiong S, Sun F, Qin G, Hu G, Wang J, Zhao L, Liang YX, Wu T et al (2018b) Repopulated microglia are solely derived from the proliferation of residual microglia after acute depletion. Nat Neurosci 21: 530 –540 Hui CW, St-Pierre A, El Hajj H, Remy Y, Hébert SS, Luheshi GN, Srivastava LK, Tremblay M (2018a) Prenatal immune challenge in mice leads to partly sex-dependent behavioral, microglial, and molecular abnormalities associated with Schizophrenia. Front Mol Neurosci 11:13 Hui CW, St-Pierre MK, Detuncq J, Aumailley L, Dubois MJ, Couture V, Skuk D, Marette A, Tremblay JP, Lebel M et al (2018b) Nonfunctional mutant Wrn protein leads to neurological deficits, neuronal stress, microglial alteration, and immune imbalance in a mouse model of Werner syndrome. Brain Behav Immun 73:450 –469 Jander S, Stoll G (1996a) Downregulation of microglial keratan sulfate proteoglycans coincident with lymphomonocytic infiltration of the rat central nervous system. Am J Pathol 148:71 –78 Jander S, Stoll G (1996b) Strain-specific expression of microglial keratan sulfate proteoglycans in the normal rat central nervous system: inverse correlation with constitutive expression of major histocompatibility complex class II antigens. Glia 18:255 –260 Jones LL, Tuszynski MH (2002) Spinal cord injury elicits expression of keratan sulfate proteoglycans by macrophages, reactive microglia, and oligodendrocyte progenitors. J Neurosci 22:4611 –4624 Joost E, Jordão MJC, Mages B, Prinz M, Bechmann I, Krueger M (2019) Microglia contribute to the glia limitans around arteries, capillaries and veins under physiological conditions, in a model of neuroinflammation and in human brain tissue. Brain Struct Funct 224:1301 –1314 Jordão MJC, Sankowski R, Brendecke SM, Sagar Locatelli G, Tai YH, Tay TL, Schramm E, Armbruster S, Hagemeyer N, Groß O et al (2019) Single-cell profiling identifies myeloid cell subsets with distinct fates during neuroinflammation. Science 363: eaat7554 Kamigaki M, Hide I, Yanase Y, Shiraki H, Harada K, Tanaka Y, Seki T, Shirafuji T, Tanaka S, Hide M et al (2016) The Toll-like receptor 4-activated neuroprotective microglia subpopulation survives via granulocyte macrophage colony-stimulating factor and JAK2/STAT5signaling. Neurochem Int 93:82 –94 Kamphuis W, Kooijman L, Schetters S, Orre M, Hol EM (2016) Transcriptional profiling of CD11c-positive microglia accumulating around amyloid plaques in a mouse model for Alzheimer’s disease. Biochim Biophys Acta 1862:1847 –1860 Kan MJ, Lee JE, Wilson JG, Everhart AL, Brown CM, Hoofnagle AN, Jansen M, Vitek MP, Gunn MD, Colton CA (2015) Arginine deprivation and immune suppression in a mouse model of Alzheimer’s disease. J Neurosci 35: 5969 –5982 Keren-Shaul H, Spinrad A, Weiner A, Matcovitch-Natan O, Dvir-Szternfeld R, Ulland TK, 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 e17 Kierdorf K, Erny D, Goldmann T, Sander V, Schulz C, Perdiguero EG, Wieghofer P, Heinrich A, Riemke P, Hölscher C et al (2013) Microglia emerge from erythromyeloid precursors via Pu.1and Irf8-dependent pathways. Nat Neurosci 16:273 –280 Kiyofuji K, Kurauchi Y, Hisatsune A, Seki T, Mishima S, Katsuki H (2015)A natural compound macelignan protects midbrain dopaminergic neurons from inflammatory degeneration via microglial arginase-1expression. Eur J Pharmacol 760:129 –135 Koellhoffer EC, McCullough LD, Ritzel RM (2017) Old maids: aging and its impact on microglia function. Int J Mol Sci 18:E769 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 e9 Lawson LJ, Perry VH, Dri P, Gordon S (1990) Heterogeneity in the distribution and morphology of microglia in the normal adult mouse brain. Neuroscience 39:151 –170 Lawson LJ, Perry VH, Gordon S (1992) Turnover of resident microglia in the normal adult mouse brain. Neuroscience 48:405 –415 Lenz KM, Nugent BM, Haliyur R, McCarthy MM (2013) Microglia are essential to masculinization of brain and behavior. J Neurosci 33:2761 –2772 Li Q, Cheng Z, Zhou L, Darmanis S, Neff NF, Okamoto J, Gulati G, Bennett ML, Sun LO, Clarke LE et al (2019) Developmental heterogeneity of microglia and brain myeloid cells revealed by deep single-cell RNA sequencing. Neuron 101:207 –223 e10 Luo J, Elwood F, Britschgi M, Villeda S, Zhang H, Ding Z, Zhu L, Alabsi H, Getachew R, Narasimhan R et al (2013) Colony-stimulating factor 1 receptor (CSF1R) signaling in injured neurons facilitates protection and survival. J Exp Med 210:157 –172 16 of 18 The EMBO Journal 38:e101997 |2019 ª2019 The Authors The EMBO Journal Vassilis Stratoulias et al Martinez FO, Gordon S (2014) The M1and M2paradigm of macrophage activation: time for reassessment. F1000Prime Rep 6:13 Mass E, Ballesteros I, Farlik M, Halbritter F, Günther P, Crozet L, Jacome-Galarza CE, Händler K, Klughammer J, Kobayashi Y et al (2016) Specification of tissue-resident macrophages during organogenesis. Science 353: aaf4238 Masuda T, Sankowski R, Staszewski O, Böttcher C, Amann L, Scheiwe C, Nessler S, Kunz P, van Loo G, Coenen VA et al (2019) Spatial and temporal heterogeneity of mouse and human microglia at single-cell resolution. Nature 566:388 –392 Matcovitch-Natan O, Winter DR, Giladi A, Vargas Aguilar S, Spinrad A, Sarrazin S, Ben-Yehuda H, David E, Zelada González 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 JZ, Manet E, Hemberg M, De Jager PL, Ransohoff RM, Regev A, Tsai LH (2017) Temporal tracking of microglia activation in neurodegeneration at single-cell resolution. Cell Rep 21: 366 –380 Matsui H, Ohgomori T, Natori T, Miyamoto K, Kusunoki S, Sakamoto K, Ishiguro N, Imagama S, Kadomatsu K (2013) Keratan sulfate expression in microglia is diminished in the spinal cord in experimental autoimmune neuritis. Cell Death Dis 4:e946 McCluskey LP, Lampson LA (2000) Local neurochemicals and site-specific immune regulation in the CNS. J Neuropathol Exp Neurol 59:177 –187 Medrano-Fernández A, Barco A (2016) Nuclear organization and 3D chromatin architecture in cognition and neuropsychiatric disorders. Mol Brain 9:83 Mildner A, Schmidt H, Nitsche M, Merkler D, Hanisch UK, Mack M, Heikenwalder M, Brück W, Priller J, Prinz M (2007) Microglia in the adult brain arise from Ly-6ChiCCR94+monocytes only under defined host conditions. Nat Neurosci 10:1544 –1553 Milior G, Lecours C, Samson L, Bisht K, Poggini S, Pagani F, Deflorio C, Lauro C, Alboni S, Limatola C et al (2016) Fractalkine receptor deficiency impairs microglial and neuronal responsiveness to chronic stress. Brain Behav Immun 55:114 –125 Mrdjen D, Pavlovic A, Hartmann FJ, Schreiner B, Utz SG, Leung BP, Lelios I, Heppner FL, Kipnis J, Merkler D et al (2018) High-dimensional single-cell mapping of central nervous system immune cells reveals distinct myeloid subsets in health, aging, and disease. Immunity 48:380 –395 e6 Nagarajan N, Jones BW, West PJ, Marc RE, Capecchi MR (2018) Corticostriatal circuit defects in Hoxb8mutant mice. Mol Psychiatry 23:1–10 Nandi S, Gokhan S, Dai XM, Wei S, Enikolopov G, Lin H, Mehler MF, Stanley ER (2012) The CSF-1receptor ligands IL-34 and CSF-1exhibit distinct developmental brain expression patterns and regulate neural progenitor cell maintenance and maturation. Dev Biol 367:100 –113 Nimmerjahn A, Kirchhoff F, Helmchen F (2005) Resting microglial cells are highly dynamic surveillants of brain parenchyma in vivo.Science 308: 1314 –1318 Okawa S, Saltó C, Ravichandran S, Yang S, Toledo EM, Arenas E, Del Sol A (2018) Transcriptional synergy as an emergent property defining cell subpopulation identity enables population shift. Nat Commun 9:2595 Olah M, Biber K, Vinet J, Boddeke HW (2011) Microglia phenotype diversity. CNS Neurol Disord Drug Targets 10:108 –118 Oosterhof N, Chang IJ, Karimiani EG, Kuil LE, Jensen DM, Daza R, Young E, Astle L, van der Linde HC, Shivaram GM et al (2019) Homozygous mutations in CSF1R cause a pediatric-onset leukoencephalopathy and can result in congenital absence of microglia. Am J Hum Genet 104: 936 –947 Orre M, Kamphuis W, Osborn LM, Jansen AHP, Kooijman L, Bossers K, Hol EM (2014) Isolation of glia from Alzheimer’s mice reveals inflammation and dysfunction. Neurobiol Aging 35:2746 –2760 Paolicelli RC, Bisht K, Tremblay M (2014) Fractalkine regulation of microglial physiology and consequences on the brain and behavior. Front Cell Neurosci 8:129 Paris I, Savage JC, Escobar L, Abiega O, Gagnon S, Hui CW, Tremblay M, Sierra A, Valero J (2018) ProMoIJ: a new tool for automatic three-dimensional analysis of microglial process motility. Glia 66:828 –845 Perdiguero EG, Klapproth K, Schulz C, Busch K, de Bruijn M, Rodewald HR, Geissmann F (2015) The origin of tissue-resident macrophages: when an erythro-myeloid progenitor is an erythro-myeloid progenitor. Immunity 43: 1023 –1024 Peterson VM, Zhang KX, Kumar N, Wong J, Li L, Wilson DC, Moore R, McClanahan TK, Sadekova S, Klappenbach JA (2017) Multiplexed quantification of proteins and transcripts in single cells. Nat Biotechnol 35: 936 –939 Pimentel-Coelho PM, Michaud JP, Rivest S (2013) Evidence for a genderspecific protective role of innate immune receptors in a model of perinatal brain injury. J Neurosci 33:11556 –11572 Ransohoff RM (2016) A polarizing question: do M1and M2microglia exist? Nat Neurosci 19:987 –991 Ravasi T, Wells C, Forest A, Underhill DM, Wainwright BJ, Aderem A, Grimmond S, Hume DA (2002) Generation of diversity in the innate immune system: macrophage heterogeneity arises from gene-autonomous transcriptional probability of individual inducible genes. J Immunol 168: 44 –50 Remington LT, Babcock AA, Zehntner SP, Owens T (2007) Microglial recruitment, activation, and proliferation in response to primary demyelination. Am J Pathol 170:1713 –1724 Réu P, Khosravi A, Bernard S, Mold JE, Salehpour M, Alkass K, Perl S, Tisdale J, Possnert G, Druid H et al (2017) The lifespan and turnover of microglia in the human brain. Cell Rep 20:779 –784 Ribeiro Xavier AL, Kress BT, Goldman SA, Lacerda de Menezes JR, Nedergaard M(2015) A distinct population of microglia supports adult neurogenesis in the subventricular zone. J Neurosci 35:11848 –11861 Río-Hortega P (1919a) El “tercer elemento de los centros nerviosos”. IV. Poder fagocitario y movilidad de la microglía. Bol Soc Esp Biol VIII: 155 –166 Río-Hortega P (1919b) El “tercer elemento”de los centros nerviosos. I. La microglía en estado normal. II. Intervención de la microglía en los procesos patológicos (células en bastoncito y cuerpos gránulo-adiposos). Bol Soc Esp Biol VIII: 69 –109 Río-Hortega P (1919c) El “tercer elemento”de los centros nerviosos. III. Naturaleza probable de la microglía. Bol Soc Esp Biol VIII: 108 –115 Río-Hortega P (1920) Estudios sobre la neuroglia. La microglía y su transformación en células en bastoncito y cuerpos gránulo-adiposos. Trab Lab Invest Biol Univ Madrid XVIII: 37 –82 Salter MW, Stevens B (2017) Microglia emerge as central players in brain disease. Nat Med 23:1018 –1027 Sato-Hashimoto M, Nozu T, Toriba R, Horikoshi A, Akaike M, Kawamoto K, Hirose A, Hayashi Y, Nagai H, Shimizu W et al (2019) Microglial SIRPa regulates the emergence of CD11c. Elife 8:e42025 Schafer DP, Lehrman EK, Kautzman AG, Koyama R, Mardinly AR, Yamasaki R, Ransohoff RM, Greenberg ME, Barres BA, Stevens B (2012) Microglia sculpt postnatal neural circuits in an activity and complement-dependent manner. Neuron 74:691 –705 Scheffel J, Regen T, Van Rossum D, Seifert S, Ribes S, Nau R, Parsa R, Harris RA, Boddeke HW, Chuang HN et al (2012) Toll-like receptor activation ª2019 The Authors The EMBO Journal 38:e101997 |2019 17 of 18 Vassilis Stratoulias et al The EMBO Journal reveals developmental reorganization and unmasks responder subsets of microglia. Glia 60:1930 –1943 Schmid CD, Sautkulis LN, Danielson PE, Cooper J, Hasel KW, Hilbush BS, Sutcliffe JG, Carson MJ (2002) Heterogeneous expression of the triggering receptor expressed on myeloid cells-2on 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 SE, Pollard JW et al (2012) A lineage of myeloid cells independent of Myb and hematopoietic stem cells. Science 336:86 –90 Schwarz JM, Sholar PW, Bilbo SD (2012) Sex differences in microglial colonization of the developing rat brain. J Neurochem 120:948 –963 Sheng J, Ruedl C, Karjalainen K (2015) Most tissue-resident macrophages except microglia are derived from fetal hematopoietic stem cells. Immunity 43:382 –393 Shigemoto-Mogami Y, Hoshikawa K, Goldman JE, Sekino Y, Sato K (2014) Microglia enhance neurogenesis and oligodendrogenesis in the early postnatal subventricular zone. J Neurosci 34:2231 –2243 Shinjo R, Imagama S, Ito Z, Ando K, Nishida Y, Ishiguro N, Kadomatsu K (2014) Keratan sulfate expression is associated with activation of a subpopulation of microglia/macrophages in Wallerian degeneration. Neurosci Lett 579:80 –85 Spittau B (2017) Aging microglia-phenotypes, functions and implications for age-related neurodegenerative diseases. Front Aging Neurosci 9:194 Squarzoni P, Oller G, Hoeffel G, Pont-Lezica L, Rostaing P, Low D, Bessis A, Ginhoux F, Garel S (2014) Microglia modulate wiring of the embryonic forebrain. Cell Rep 8:1271 –1279 Stevens B, Allen NJ, Vazquez LE, Howell GR, Christopherson KS, Nouri N, Micheva KD, Mehalow AK, Huberman AD, Stafford B et al (2007) The classical complement cascade mediates CNS synapse elimination. Cell 131: 1164 –1178 Stowell RD, Wong EL, Batchelor HN, Mendes MS, Lamantia CE, Whitelaw BS, Majewska AK (2018) Cerebellar microglia are dynamically unique and survey Purkinje neurons in vivo.Dev Neurobiol 78:627 –644 Streit WJ (2002) Microglia as neuroprotective, immunocompetent cells of the CNS. Glia 40:133 –139 Tay TL, Béchade C, D’Andrea I, St-Pierre MK, Henry MS, Roumier A, Tremblay ME (2017a) Microglia gone rogue: impacts on psychiatric disorders across the lifespan. Front Mol Neurosci 10:421 Tay TL, Mai D, Dautzenberg J, Fernández-Klett F, Lin G, Sagar Datta M, Drougard A, Stempfl T, Ardura-Fabregat A, Staszewski O et al (2017b) A new fate mapping system reveals context-dependent random or clonal expansion of microglia. Nat Neurosci 20:793 –803 Tay TL, Savage JC, Hui CW, Bisht K, Tremblay M (2017c) Microglia across the lifespan: from origin to function in brain development, plasticity and cognition. J Physiol 595:1929 –1945 Tay TL, Sagar DJ, Grün D, Prinz M (2018) Unique microglia recovery population revealed by single-cell RNAseq following neurodegeneration. Acta Neuropathol Commun 6:87 Trapnell C (2015) Defining cell types and states with single-cell genomics. Genome Res 25:1491 –1498 Tremblay M, Lowery RL, Majewska AK (2010) Microglial interactions with synapses are modulated by visual experience. PLoS Biol 8:e1000527 Tremblay M, Zettel ML, Ison JR, Allen PD, Majewska AK (2012) Effects of aging and sensory loss on glial cells in mouse visual and auditory cortices. Glia 60:541 –558 Tremblay M, Lecours C, Samson L, Sánchez-Zafra V, Sierra A (2015) From the Cajal alumni Achúcarro and Río-Hortega to the rediscovery of neverresting microglia. Front Neuroanat 9:45 Walz W, Lang MK (1998) Immunocytochemical evidence for a distinct GFAPnegative subpopulation of astrocytes in the adult rat hippocampus. Neurosci Lett 257:127 –130 Wang Y, Berezovska O, Fedoroff S (1999) Expression of colony stimulating factor-1receptor (CSF-1R) by CNS neurons in mice. J Neurosci Res 57: 616 –632 Wang Y, Szretter KJ, Vermi W, Gilfillan S, Rossini C, Cella M, Barrow AD, Diamond MS, 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 Weil ZM, Norman GJ, DeVries AC, Nelson RJ (2008) The injured nervous system: a Darwinian perspective. Prog Neurobiol 86:48 –59 Wendeln AC, Degenhardt K, Kaurani L, Gertig M, Ulas T, Jain G, Wagner J, Häsler LM, Wild K, Skodras A et al (2018) Innate immune memory in the brain shapes neurological disease hallmarks. Nature 556:332 –338 Wlodarczyk A, Løbner M, Cédile O, Owens T (2014) Comparison of microglia and infiltrating CD11c + cells as antigen presenting cells for T cell proliferation and cytokine response. J Neuroinflammation 11:57 Wlodarczyk A, Cédile O, Jensen KN, Jasson A, Mony JT, Khorooshi R, Owens T (2015) Pathologic and protective roles for microglial subsets and bone marrowand blood-derived myeloid cells in central nervous system inflammation. Front Immunol 6:463 Wlodarczyk A, Holtman IR, Krueger M, Yogev N, Bruttger J, Khorooshi R, Benmamar-Badel A, de Boer-Bergsma JJ, Martin NA, Karram K et al (2017) A novel microglial subset plays a key role in myelinogenesis in developing brain. EMBO J 36:3292 –3308 Wlodarczyk A, Benmamar-Badel A, Cédile O, Jensen KN, Kramer I, Elsborg NB, Owens T (2018) CSF1R stimulation promotes increased neuroprotection by CD11c+microglia in EAE. Front Cell Neurosci 12:523 Wogram E, Wendt S, Matyash M, Pivneva T, Draguhn A, Kettenmann H (2016) Satellite microglia show spontaneous electrical activity that is uncorrelated with activity of the attached neuron. Eur J Neurosci 43: 1523 –1534 Xavier AL, Lima FR, Nedergaard M, Menezes JR (2015) Ontogeny of CX3CR151EGFP expressing cells unveil microglia as an integral component of the postnatal subventricular zone. Front Cell Neurosci 9:37 Xu J, Zhu L, He S, Wu Y, Jin W, Yu T, Qu JY, Wen Z (2015) Temporal-spatial resolution fate mapping reveals distinct origins for embryonic and adult microglia in zebrafish. Dev Cell 34:632 –641 Yeh FL, Hansen DV, Sheng M (2017) TREM2, microglia, and neurodegenerative diseases. Trends Mol Med 23:512 –533 Zeisel A, Muñoz-Manchado AB, Codeluppi S, Lönnerberg P, La Manno G, Juréus 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 Zhang C, Lam TT, Tso MO (2005) Heterogeneous populations of microglia/ macrophages in the retina and their activation after retinal ischemia and reperfusion injury. Exp Eye Res 81:700 –709 Zhang H, Muramatsu T, Murase A, Yuasa S, Uchimura K, Kadomatsu K (2006) N-Acetylglucosamine 6-O-sulfotransferase-1is required for brain keratan sulfate biosynthesis and glial scar formation after brain injury. Glycobiology 16:702 –710 License: This is an open access article under the terms of the Creative Commons Attribution 4.0 License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. 18 of 18 The EMBO Journal 38:e101997 |2019 ª2019 The Authors The EMBO Journal Vassilis Stratoulias et al