Microglia present unique transcriptional signatures related to their activity and the environmental context. Under homeostatic conditions, certain surface proteins are highly expressed, such as the CX3-motif chemokine receptor (CX3CR1), the P2Y12 purinergic receptor (P2RY12), and transmembrane protein 119 (TMEM119), among others (Song and Colonna, 2018; Paolicelli et al., 2022). However, microglial states induced by pathological insults, such as the DAM and microglial neurodegenerative phenotype (MGnD), adopt a distinct gene signature characterized by downregulation of homeostatic Cx3cr1, P2ry12, and Tmem119, together with upregulation of Trem2, Apoe, Clec7a, and Cd11c (Keren-Shaul et al., 2017; Krasemann et al., 2017; Butovsky and Weiner, 2018; Deczkowska et al., 2018; Paolicelli et al., 2022). While their functions have not been fully elucidated, DAM are involved in phagocytosis, clearance of apoptotic cell bodies and inflammatory responses (Deczkowska et al., 2018; Dubbelaar et al., 2018; Casali and Reed-Geaghan, 2021). The case of DAM illustrates that certain markers can differentially identify key microglial states during homeostasis and pathology, and these will be the subject of the following discussion. These markers are often used in immunostaining but also are frequently employed in Cre/lox systems, knockout models, using FACS and other methods (Ruan et al., 2020; Perrone et al., 2023). As an illustration of the diversity of techniques used to study microglial markers, Li et al. (2019b) combined FACS, MACS, flow cytometry and q-PCR on whole adult mouse brain tissue to investigate differences between microglial and macrophage responses to intracerebral hemorrhage. Their findings revealed elevated expressions of Tmem119 and Sall1 in the sorted microglial cell population (Ly6g^–CD11b⁺CD45^intPI^–), while Ccr2 and CD69 were highly expressed in the monocyte/macrophage sorted cell population (Ly6g^–CD11b⁺CD45^highPI^–), highlighting the utility of multifaceted approaches using multiple markers to reliably distinguish microglia. Notably, microglia expressing both Tmem119 and Sall1 exhibited increased expression following intracerebral hemorrhage, whereas monocytes/macrophages maintained consistent levels of both Ccr2⁺/CD69⁺ in the sham and hemorrhage groups (Li et al., 2019b). As discussed in the following sections, distinct sets of markers are commonly used to identify microglia depending on which technique is employed.

CD11b, one of the first identified β-glucan receptors, is an abundant surface protein primarily expressed on cells of the myeloid lineage that has been reliably used for microglial visualization. CD11b assists in endothelial adhesion and regulates the uptake of complement-coated particles by phagocytosis (Xia et al., 1999; Nikodemova and Watters, 2012; Schmid et al., 2018). As CD11b is also expressed by peripherally-derived myeloid cells in the CNS (Fischer and Reichmann, 2001; Greter et al., 2015), it must be paired together with other markers to identify microglia more specifically. It is often co-utilized with universally expressed leukocyte surface proteins such as CD45, the combination of which has utility for identifying microglia by flow cytometry (Bruzelius et al., 2021).

CD45, present on all leukocytes in varying levels, and CD68, a lysosomal marker present in endosomes and lysosomes of the mononuclear phagocytic lineages [including microglia, border-associated macrophages (BAMs) and macrophages], can be used to identify microglia in the CNS and other immune cells (Hendrickx et al., 2017; Rangaraju et al., 2018; Waller et al., 2019; Bruzelius et al., 2021). CD45, a protein tyrosine phosphatase, plays a role in the proliferation and differentiation of immune cells (Penninger et al., 2001), but tends to be expressed at lower levels on microglia than on CNS-infiltrating myeloid cells during non-inflammatory conditions, while CD68 is associated with inflammatory responses, phagolysosomal activity, and the regulation of antigen presentation (Song et al., 2011; Chistiakov et al., 2017). Both CD45 and CD68 contribute to the microglial sensome under homeostatic conditions, although CD45⁺ and CD68⁺ signatures are associated with microglial reactivity in response to inflammation, trauma or pathology (Perego et al., 2011; Hoogland et al., 2015; Hendrickx et al., 2017; Rice et al., 2017; Swanson et al., 2023). In combination with other markers, such as ionized calcium-binding adaptor molecule 1 (IBA1) and CD11b, CD45 can aid in distinguishing microglia from other CNS infiltrated myeloid cells and non-parenchymal cell populations, using, for instance, flow cytometry (Waller et al., 2019; Bruzelius et al., 2021). Microglia are generally identifiable as CD11b⁺CD45^int, in contrast to CD11b⁺CD45^hi, being usually attributed to CNS-infiltrating myeloid cells or BAMs at a steady state (Greter et al., 2015; Martin et al., 2017; Rubino et al., 2018; Honarpisheh et al., 2020). Nevertheless, under specific conditions, microglia may exhibit CD45^high expression. Studies conducted in models of stroke, cerebral amyloid angiopathy, and aging indicate that Tmem119 and P2RY12 expression is present in both CD45^int and CD45^high FACS-sorted cells myeloid cell populations (Honarpisheh et al., 2020). Moreover, a small population of CD11b⁺CD45^high mononuclear phagocytes has been identified by flow cytometry in the surroundings of Aβ plaques in the 5XFAD mouse model of Alzheimer’s disease (AD) pathology, and these cells exhibit transcriptional similarities to DAM, such as upregulation of Trem2 and Cd11c (Rangaraju et al., 2018). Additionally, CD44 is also often employed to emphasize CNS-infiltrated myeloid cells, given that macrophages, for example, typically exhibit CD44^high expression. In pathology, microglia can exhibit CD44 expression, especially when exposed to lipopolysaccharide (LPS) and in degenerative states like amyotrophic lateral sclerosis. The findings also indicate that primary cultured microglial cells express CD44 upon exposure to LPS and interferon-γ. Moreover, the results align with increased CD44 expression observed following treatment with interferon-γ and TNF-α (Matsumoto et al., 2012; Liu S. et al., 2023).

Thus, the identification of microglia by flow cytometry in pathological conditions remains complicated, requiring fate mapping approaches to distinguish these lineages robustly. Additionally, distinguishing microglia from CNS-infiltrated macrophages is crucial for the precise characterization and understanding of the functionally of each cell type. Notably, BAMs form a distinct, specialized population naturally located in the choroid plexus (CP), meninges and perivascular spaces (Dermitzakis et al., 2023). Given the immune-privileged nature of the CNS, it is naturally constituted and protected by microglia, parenchymal tissue-resident macrophage, along with non-parenchymal cells, the BAMs. The resemblance to microglia is significant, as these cells are also derived from the yolk sac and share certain commonly expressed genes with microglia. For instance, they exhibit low expression levels of P2y12 and Tmem119, while microglia express these genes at higher levels (Rajan et al., 2020). Single-cell approaches, using mouse enriched subdural meninges (SDM) CP^epi-BAMs, have also demonstrated that BAM subsets express common DAM signature genes, including an upregulation of genes involved in lipid metabolism or phagocytosis, such as Cst7, Apoe, Clec7a, and Lpl (Van Hove et al., 2019). BAMs often express lymphatic-like gene signatures, such as Pdpn and Lyve1, frequently associated with cell migration. Expression levels naturally differ among distinct subpopulations of BAMs located in various non-parenchymal regions, such as the or meninges (Møllgård et al., 2023). Interestingly, microglia and BAMs share the same erythromyeloid progenitors (EMPs) in the yolk sac, however, fate mapping approaches revealed that these EMPs seem to depend on CD206 expression for their differentiation into microglia or BAMs (Masuda et al., 2022). While only a few microglia start to express CD206 at the embryonic day (E)12.5, BAMs start to highly express it at E10.5 (Masuda et al., 2022). BAMs will be further discussed in the subsequent fate mapping sections. To ensure the characterization and distinction of microglia, BAMs and peripherally derived myeloid cells, complementary techniques are often employed, such as flow cytometry, fate mapping and single-cell transcriptome analysis. IHC can also aid in discerning microglia from peripherally-derived myeloid cells, for example based on higher levels of IBA1 and TMEM119 expression in microglia (Kishimoto et al., 2019).

Both CD86 and CD206 characterize reactive microglia but are associated with distinct functional properties and microglial phenotypes, which are extensively examined in immunohistochemistry and flow cytometry. The C-type lectin CD206 plays a role in recognizing pathogens and clearing glycoproteins from the circulation (Suzuki et al., 2018; Tanaka et al., 2021). Not specific to microglia, CD86 and CD206 are also expressed by other cells in the CNS, such as astrocytes and BAMs, and throughout the body, being found in dendritic cells, macrophages/monocytes and B/T-lymphocytes (Xing et al., 2018; Sun et al., 2019; Trzupek et al., 2020). Upregulation of CD206 is observed in pathological conditions, such as glioma, where increased microglial pinocytosis and phagocytosis mediate the clearance of debris and resolution of inflammation (Tanaka et al., 2021). CD86, also a transmembrane protein, provides a costimulatory signal essential for the activation and proliferation of T cells. Hence it is abundantly expressed on antigen-presenting cells, including microglia, macrophages, dendritic cells, and B cells, and is upregulated in response to injury (Louveau et al., 2015; Hellström Erkenstam et al., 2016). Remarkably, it is seen an increase in both CD86⁺ and CD206⁺ subsets among brain CD11b⁺ cells, in models of traumatic brain injury, neonatal brain hypoxia-ischemia and cerebral ischemia-reperfusion (Jin et al., 2012; Hellström Erkenstam et al., 2016; Bai et al., 2017; Yusuying et al., 2022).

As microglia respond to many cues in the microenvironment, they also naturally express receptors associated with neuronal communication (Butler et al., 2021). CX3CR1 is widely expressed in the CNS, for instance, in macrophages and microglia, and its ligand, fractalkine (CX3CL1), is an important chemokine produced by neurons (Wolf et al., 2013; Szepesi et al., 2018). The CX3CL1-CX3CR1 axis mediates the development and plasticity of neuronal circuits, with functional consequences for brain connectivity, as notably shown by comparing the CX3CR1^KO/KO mice with CX3CR1^KO/+ mice (Drissen et al., 2019; Lauro et al., 2019). Under physiological conditions, CX3CR1-expressing microglia in the brain are extensively involved in dynamic surveillance and maintenance of brain homeostasis (Lauro et al., 2019). CX3CR1 signaling is also important for microglial motility and the regulation of the microglial response to injury and inflammation. Studies employing immunostaining and q-PCR in mouse models of AD pathology and ischemic stroke have shown that CX3CR1^high microglia have increased migration to sites of pathology, whereas CX3CR1^low microglia tend to remain in a surveillance state (Tang et al., 2014; Hickman et al., 2019). Similarly, CX3CR1-deficient microglia exhibit delayed migration toward sites of laser-induced injury and impaired repopulation following depletion (Liang et al., 2009; Zhang et al., 2018).

In recent years, TMEM119 had emerged as a microglial marker with high specificity (Bennett et al., 2016; Satoh et al., 2016). Initially identified as a regulator of osteoblast differentiation, TMEM119 is thought to regulate the Wnt/β-catenin pathway, important for proliferation, migration, and genetic stability, but further research is required to fully elucidate its functions in microglia (Satoh et al., 2016; Yang et al., 2021). Selectively expressed by yolk sac-derived cells in the CNS (e.g., microglia and BAMs in the CNS), TMEM119 broadly distinguishes resident microglia from infiltrating macrophages (Haynes et al., 2006; Segal and Giger, 2016; Bohnert et al., 2020; Vankriekelsvenne et al., 2022). TMEM119 immunostaining can be combined with IBA1 to permit distinct visualization of microglia and brain-infiltrating macrophages by immunofluorescence (van Wageningen et al., 2019). However, caution should be exercised when using TMEM119 as a selective marker for microglia. TMEM119 protein expression has been shown to be state-dependent, and can be strongly downregulated by reactive microglia in pathological conditions (Haynes et al., 2006; Segal and Giger, 2016; Bohnert et al., 2020; Vankriekelsvenne et al., 2022).

P2Y receptors are metabotropic purinergic receptors that signal in response to adenosine triphosphate (ATP), notably released by injured cells (van Wageningen et al., 2019; von Kügelgen, 2019; Gómez Morillas et al., 2021). These receptors are extensively present in microglia, but also detected in astrocytes within the CNS, in both humans and rodents (Franke et al., 2004; Milior et al., 2020; Walker et al., 2020). P2RY12 is particularly important for microglia-neuron communication, particularly through regulating microglial chemotaxis toward neuronal-derived ATP in response to injury (Haynes et al., 2006). P2RY12 serves as an effective marker for observing microglial motility, frequently revealing surveillance movements, membrane ruffling, and live imaging of process extension or retraction. In fact, in vitro and in vivo studies have shown that microglia lacking P2RY12 receptors are unable to migrate or extend processes toward nucleotides (Haynes et al., 2006; Lin et al., 2020). Microglial P2RY12 is also a key mediator of synaptic plasticity and behavior (Sipe et al., 2016; Lowery et al., 2021) and contributes to synaptic loss in models of chronic stress (Bollinger et al., 2023). Similarly to TMEM119, P2RY12 is often downregulated under inflammatory or pathological conditions (Haynes et al., 2006; van Wageningen et al., 2019; Lin et al., 2020). In post-mortem tissue samples from multiple sclerosis patients, immunoreactivity for both TMEM119 and P2RY12 was decreased in the centers of white matter lesions, which correlated with the presence of lymphocyte infiltrates and proinflammatory cytokines (van Wageningen et al., 2019). However, in a mouse model of status epilepticus, microglial P2RY12 expression and purinergic signaling were increased in the hippocampus (Avignone et al., 2008), highlighting the complex dynamics of P2RY12 expression across pathologies. Likewise, the pattern of purine-induced motility in human microglia within epileptic tissue appears to have some similarities to that observed in rodent microglia. As reported by Milior and colleagues, low doses of purine in human tissue promoted microglial branch extension, whereas in high concentrations, process retraction and membrane ruffling were facilitated (Milior et al., 2020). Notably, in the context of microglial branch extension, activation of P2Y12 was essential, mirroring that observed in rodents. Yet, the mechanism for microglial processes retraction was mediated by the activation of P2Y1 and P2Y13 receptors, rather than A2A—as classically seen in rodents (Orr et al., 2009; Milior et al., 2020). P2RY12 is extensively used in KO models, fate mapping, and reporter models driven by the purinergic signaling promoter, common within heterozygous GFP reporter mice under the fractalkine receptor promoter or TMEM119. This combination also enables in vivo imaging using two-photon microscopy (Bisht et al., 2021).

Intracellular proteins are important for vital microglial functions, such as cell organization, metabolism and adhesion, and can consistently aid in microglial visualization (Zhao et al., 2022). IBA1 is the most common cytoplasmic protein used for microglial immunostaining (Ohsawa et al., 2004; Tremblay et al., 2010). It was first isolated from rats in 1996, and labels both microglia, macrophages/monocytes and BAMs in the CNS (Imai et al., 1996; Sasaki et al., 2001; Jurga et al., 2020). IBA1 is primarily responsible for cytoskeletal reorganization, which enables processes such as membrane ruffling, migration and phagocytosis (Sasaki et al., 2001; Ohsawa et al., 2004; Hopperton et al., 2018; Sobue et al., 2021). IBA1 is commonly used in IHC, mostly in conjugation with other antibodies of interest such as TMEM119 and P2RY12, to aid the distinction of infiltrating myeloid cells, which stain dimly with these antibodies. IBA1-IHC can also be coupled to other techniques, such as CX3CR1-GFP reporter mice or single-cell approaches to describe diverse microglial subpopulations across various physiological and pathological conditions in the CNS (Ibanez et al., 2019; Lituma et al., 2021; González Ibáñez et al., 2023). In both white and gray matter, IBA1 has high sensitivity for microglia and appears to be expressed across a majority of microglial states (Shapiro et al., 2009). Moreover, IBA1 expression is found across many vertebrate and invertebrate species, highlighting its expression in humans and rodents (Swanson et al., 2020; Sharma et al., 2021). Interestingly, studies show that IBA1 expression in human microglia may be altered in pathological states, including AD (Hopperton et al., 2018). Under these conditions, microglial morphology and ultrastructure can aid in identifying microglial phenotypes. For instance, dark microglia, a phenotype identified by EM and found in mouse models and human post-mortem samples of AD pathology, express very low levels of IBA1 but can be distinguished by their unique ultrastructural characteristics, notably an electron-dense (dark) cytoplasm and nucleoplasm (Bisht et al., 2016). In contrast to mouse models, dark microglia observed to date in human patients with AD stain positive for IBA1 (St-Pierre et al., 2022a). Dystrophic microglia represent another morphologically-defined phenotype, staining positive for IBA1 but exhibiting thin, fragmented processes (Streit et al., 2004, 2009), and are found in brains of human AD patients in proportionately greater quantities than during normal aging (Shahidehpour et al., 2021). These microglia are thought to be senescent and dysfunctional, and their presence correlates with neuropathology in AD (Streit et al., 2009, 2020).

The advent of novel transcriptomic techniques, such as microarrays and RNA sequencing, has facilitated the identification of key microglial regulators of transcription (Olah et al., 2020). Researchers recently demonstrated that the Spalt-like 1 gene (Sall1) and the hexosaminidase β-subunit (Hexb) are selectively expressed in murine and human brain microglia and play important roles in normal microglial maturation and function (Butovsky and Weiner, 2018). In mice, SALL1 is expressed by microglia and by other mononuclear or resident glial cells of the CNS, such as astrocytes (Buttgereit et al., 2016; Chi et al., 2019). SALL1 is highly expressed in young murine CNS microglia associated with critical functions, such as neural maturation and synaptic pruning (Buttgereit et al., 2016; Holtman et al., 2017; Sobue et al., 2021; Scott et al., 2022). Microglia-specific Sall1 deletion in vivo results in altered morphology and converts surveillant microglia to a reactive, proinflammatory phenotype (Buttgereit et al., 2016). Additionally, SALL1 is expressed during development and persists throughout the life of the cell, making it a useful marker for tracking microglia across the lifespan (Buttgereit et al., 2016).

Similarly, Hexb is a lysosomal enzyme that plays critical role in lysis processes, breaking down fatty compounds such as sphingolipids and gangliosides (Kuil et al., 2019). Recent studies using RNAscope and single-cell RNA sequencing have shown that Hexb is expressed in brain microglia and its expression is highly restricted to these cells within the CNS (Masuda et al., 2020; Faust et al., 2023). Notably, it was not detected in CNS-associated macrophages according to Masuda et al. (2020); however, it has been identified in non-CNS cells, such as Ly6C⁺ monocytes obtained from the bone marrow (Hickman et al., 2013; Monaghan et al., 2019; Masuda et al., 2020; Shah et al., 2022). HEXB may also play a role in pathology, with elevated gene expression of HEXB associated with poor prognosis in human glioblastoma (Jia et al., 2021). Although HEXB immunostaining has been utilized in combination with TMEM119 to directly identify microglia (Jia et al., 2021), Hexb-based gene reporter and fate mapping tools, have also recently provided important insights into the behavior and functions of microglia in vivo (Masuda et al., 2020). Additionally, the promoter region of Hexb has recently been characterized, enabling the study of Hexb as a potential target for microglia-targeted gene therapies in humans (Shah et al., 2022).

Classically, transcription factors modulate transcription within the nucleus, providing effective targets for identifying microglial cells. For instance, PU.1 is exclusively found in microglia in the brain, however, it is also expressed throughout the body by other cells, such as macrophages, granulocytes, and B cells (Pimenova et al., 2021; Cakir et al., 2022). PU.1 is a crucial regulator of microglial development and homeostatic function. In fact, KO of PU.1 in BV2, an immortalized mouse microglial cell line, abates most of the homeostatic microglial signatures. Conversely, it upregulates genes associated with DAMs, including Lpl and Clec7a, as assessed through q-PCR and Western Blotting in BV2-lineage cells (Pimenova et al., 2021). Moreover, elevated PU.1 expression is associated with an elevated risk of AD, whereas reduced expression is believed to offer some protective effect and contribute to improving inflammation balance by microglia, result observed in primary human mixed glial cultures following siRNA-mediated KO of PU.1 (Rustenhoven et al., 2018; Pimenova et al., 2021; Ralvenius et al., 2023). Sall1 can also be referred to as a nuclear marker for microglia. In mice, the initiation of Sall1 expression occurs at E11 to 12 in hematopoietic progenitor cells derived from the yolk sac. These cells infiltrate the developing brain and ultimately differentiate into resident microglia. Notably, the expression of Sall1 is dependent on TGFβ1 signaling, an instrumental pathway in microglial differentiation and survival (Fixsen et al., 2023). The expression of Sall1 by microglia is context-dependent, it can decline when microglia are transplanted from the brain to an in vitro environment, underscoring the importance of the brain micro-environment to sustain an authentic in vivo microglial phenotype (Gosselin et al., 2017; Fixsen et al., 2023).

Similarly, single-cell technology has identified ZEB1 and MAFB as reliable nuclear markers for microglia. MAFB exhibits diffuse expression throughout the brain and body in both mice and humans, including neurons, microglia/macrophages, other glial cells, Kupffer cells in the liver, and Hofbauer cells, fetal origin macrophages in placenta (Anderson et al., 2023; Liu S. et al., 2023; Liu W. et al., 2023; Vondra et al., 2023). On the other hand, ZEB1 is expressed in radial glial cells and intermediate progenitor cells, as well as mature astrocytes, but seems to be downregulated in neuronal lineages (Gupta et al., 2021). Using multiomics and a methodology to delineate peak-gene-transcription factor trios in nuclei isolated from human post-mortem DLPFC tissues, Anderson et al. (2023) showed that the regulation of MAFB in microglia has been linked to high DAM-like signatures, including TRL3 and CD84. Interestingly, ZEB1 in neurons was identified in half of all AD-specific trios with target genes involved in regulating ion channel signaling (ITPR1, CAMK2A, CACNB3). In contrast, ZEB1 was not identified in control (non-AD) specific trios, suggesting that it may experience upregulation under pathological conditions (Maimaiti et al., 2019; Poonaki et al., 2022; Anderson et al., 2023).

The extracellular matrix (ECM) is a non-cellular component present in all tissues that provides a physical scaffold, biochemical and biomechanical cues, and helps organize cellular morphology and structure. Each tissue has unique ECM characteristics and structure, with specific growth factors and enzymes determining its composition and function (Theocharis et al., 2016; Lam et al., 2019). As an extracellular component, the ECM is constituted of a variety of matrix proteins that are produced by many different cell types. Components of the ECM can actively influence the adhesion, migration, and morphology of microglial cells. For instance, various components of the ECM, such as fibronectin, vitronectin and laminin, interact with microglial cell surface receptors. For instance, microglia can adhere to these specific ECM molecules through integrin receptors, reshaping microglial activity and motility according to the ECM environment (Milner and Campbell, 2002). Notably, neuronal IL 33 contributes significantly to synapse plasticity in the hippocampus of adult mice. While IL-33 guides microglial engulfment of the ECM, its deficiency results in compromised ECM engulfment, leading to the accumulation of ECM proteins in proximity to synapses (Nguyen et al., 2020). Notably, all synapses are surrounded by interstitial extracellular matrix (iECM), mostly regulating the synapse volume, however, certain synapses are enclosed by a condensed version of the ECM, known as perineuronal nets (PNNs) (Fawcett et al., 2022). Specifically, PNNs form a specialized ECM type in the CNS, providing structural support for cellular elements and maintaining synapse homeostasis. These nets are mostly formed by chondroitin sulfate proteoglycans (CSPGs) attached to a hyaluronan structure (Fawcett et al., 2022). PNNs form a net-like structure that surrounds the cell body and dendrites, enveloping synapses. It is now evident that PNNs significantly contribute to plasticity control. Interestingly, approximately 80% of parvalbumin⁺ neurons in the primary somatosensory cortex are coated with PNNs. Venturino et al. (2021) showed that microglia closely interact with parvalbumin neurons in the cortex using immunostaining against IBA1 and parvalbumin itself in adult mice. Moreover, parvalbumin neurons and IBA1⁺ microglia significantly increased matrix metalloproteinase (MMP)-9 levels upon ketamine exposure in cortical layers 3–5 (Venturino et al., 2021).

Specifically, microglia, as well as monocytes, macrophages, astrocytes and lymphocytes, produce MMPs such as MMP-3 and MMP-9, which mediate the degradation of multiple ECM components (Lindberg et al., 2001; Patnaik et al., 2021). MMPs play important physiological roles in synaptic plasticity, learning and memory (Meighan et al., 2006, 2007; Szepesi et al., 2014; Bijata et al., 2017). In fact, MMP-9 is especially vital for the formation of sensory circuits in the early postnatal period, demonstrating increased activity that aligns with synaptic reorganization during critical period plasticity and developmental windows across the CNS (Hanania et al., 2012; Reinhard et al., 2015; Shih et al., 2021; Akol et al., 2022). MMPs have been implicated in various disease states, with hippocampal induction of MMP-9 activity recently shown to mediate the development of depressive-like behavior in a mouse model of chronic stress (Bijata et al., 2022). Microglial upregulation of MMPs is also associated with inflammation or injury (Rosell et al., 2006; Colognato and Tzvetanova, 2011; Könnecke and Bechmann, 2013). Microglial-secreted MMP-3 and MMP-9 in turn influence microglia themselves, promoting production of reactive oxygen species and proinflammatory cytokines such as TNF-α and IL-1β (Kim et al., 2005; Woo et al., 2008; Könnecke and Bechmann, 2013). Recently, Kim J. et al. (2021) employed MMP-9 staining by IHC to identify microglia in the aqueous humor of patients with age-related macular degeneration and in related mouse models of choroidal neovascularization. MMP-9 expression co-localized with IBA1⁺ reactive microglia defined by an amoeboid morphology, and choroidal lesions. Notably, MMP-9 expression was suppressed by minocycline, a modulator of microglia (Kim J. et al., 2021). Thus, while the ECM itself is not specific to microglia, the use of specific ECM proteins as markers in combination with other techniques can provide a means of identifying and tracking reactive microglial states throughout physiological and pathological contexts.

The advent of homologous recombination-based techniques has allowed researchers to place a reporter gene of interest into a microglia-specific locus, replacing the endogenously expressed allele. This has been used effectively to create mouse models expressing fluorescent reporters in place of myeloid or microglia-specific proteins such as CX3CR1 (Jung et al., 2000). Constitutive reporter models have proven immeasurably useful for the visualization, identification, and study of microglia through the targeting of countless microglial genes such as Cx3cr1 (Garcia et al., 2013), Tmem119 (McKinsey et al., 2020; Ruan et al., 2020), Sall1 (Buttgereit et al., 2016; Baba et al., 2019) and Hexb (Masuda et al., 2020). The intrinsic fluorescence of reporter molecules such as eGFP and tdTomato (tdT) eliminates the need for IHC and other techniques requiring cell fixation, allowing researchers to gain insights into microglial interactions with their surroundings in an in vivo context. However, a disadvantage to this approach is the potential presence of mutations incurred by the random insertion of the construct into the genome (Wieghofer et al., 2015).

Modern approaches extensively employ the Cre/loxP system, which allows cell type-specific gene inactivation or activation (Orban et al., 1992). In this paradigm, the bacteriophage-derived Cre recombinase is expressed under cell-specific control, and precisely and irreversibly excises a distal sequence of DNA flanked by upstream and downstream loxP sequences, known as the “floxed” sequence (Orban et al., 1992; Wieghofer et al., 2015). Cre can be expressed in a microglial locus, for instance, under the Cx3cr1 promoter (Parkhurst et al., 2013; Yona et al., 2013), in mice engineered to harbor a construct consisting of a strong constitutive promoter, often Rosa26, followed by a floxed STOP codon preceding a reporter gene such as YFP (abbreviated R26^YFP) (Friedrich and Soriano, 1991; Wieghofer et al., 2015). In this example, Cre expression is governed by the endogenous regulation of Cx3cr1 and is therefore constitutively expressed in microglia and other myeloid cells that normally would highly express Cx3cr1 (Parkhurst et al., 2013; Yona et al., 2013). In CX3CR1⁺ cells, Cre excises the floxed STOP codon, leading to constitutive expression of the reporter under the control of the strong promoter. Hence, CX3CR1⁺ cells are labeled with the reporter gene of interest, enabling, for example, their identification via fluorescence microscopy or flow cytometry, or FACS-mediated isolation. Despite their utility, constitutive reporter models are limited by their property of labeling all the cells that express the gene of interest at any point during development, as Cre-mediated excision of the floxed STOP codon is irreversible. This leads to off-target cell labeling, potentially confounding results. For example, neurons transiently express Cx3cr1 during development and are labeled in Cx3cr1^Cre-based constitutive models (Haimon et al., 2018; Zhao et al., 2019), highlighting the need to interpret the results of such models cautiously. This issue is resolved with inducible reporter systems, discussed below. Reporter models to study microglia are summarized in Table 2.

More recently, the advent of inducible Cre-based systems has revolutionized the field, enabling precise temporal labeling of specific cell populations and their progeny. This approach makes use of a modified estrogen receptor (ER, MER, ERT, or ERT2), which can bind the estrogen analog 4-hydroxytamoxifen but not endogenous estrogens (Littlewood et al., 1995; Metzger et al., 1995; Feil et al., 1997; Indra et al., 1999). Development of CreER or CreERT2 fusion proteins has enabled researchers to control when Cre expression is induced, thus allowing temporal-specific induction of reporter gene expression (Feil et al., 1997; Hayashi and McMahon, 2002). Since Cre-mediated excision of the floxed sequence is irreversible, reporter gene expression is maintained in all subsequent daughter cells following mitosis, enabling the tracking of specific cell populations and their progeny (Hayashi and McMahon, 2002; Wieghofer et al., 2015). Drawing on our earlier example of Cx3cr1, placing CreER under the control of Cx3cr1 (Cx3cr1^CreER) and administering tamoxifen at a specific time in the mouse lifespan would induce constitutive reporter gene expression in all cells expressing Cx3cr1 at that timepoint, as well as in all progenies of those cells. This critical property of inducible Cre systems has allowed researchers to selectively deplete microglia, to precisely map microglial populations in a spatiotemporally controlled manner, and to identify key activities of microglia during development, homeostasis, and pathology (Zhao et al., 2019; Dai et al., 2022; Sahasrabuddhe and Ghosh, 2022), as discussed in the following sections.

However, it is important to note that inducible microglia-targeted CreER lines vary considerably in their stability, with a recent analysis showing that Cx3cr1^CreER tools demonstrate particularly high rates of spontaneous recombination in the absence of tamoxifen, compared to Tmem119^CreER- or Hexb^CreER-based lines (Faust et al., 2023). The authors also demonstrated, in some cases, reduced expression of the endogenous gene under the promoter driving CreER expression (Faust et al., 2023); as the genes used often encode essential microglial proteins, this could impair the function of Cre-targeted microglia and confound results. Although microglial gene expression is largely unaffected by tamoxifen, it was recently demonstrated that tamoxifen administration can itself lead to lower protein expression of P2RY12 (Faust et al., 2023). These issues represent important considerations when interpreting data from Cre-based microglial reporter, KO, and fate-mapping models.

Cre-based reporter systems are inherently limited by the specificity of the target gene for the cell type of interest. For example, inducible Cx3cr1-based systems, commonly used to target microglia, also affect CNS-associated and perivascular macrophages (Haimon et al., 2018; Masuda et al., 2020; Faust et al., 2023). To overcome this limitation, binary “split Cre” transgenic systems have been developed, requiring the cell of interest to simultaneously express two separate fragments (NCre and CCre) of the Cre recombinase enzyme placed under the control of different cell-specific promoters, which dimerize into a functional protein only when co-expressed (Jullien et al., 2003; Hirrlinger et al., 2009). In the context of microglia, this approach was leveraged using the clustered regularly interspaced short palindromic repeats (CRISPR)/Cas9 system to co-express NCre and CCre under the control of Cx3cr1 and Sall1, selectively targeting Cx3cr1⁺/Sall1⁺ microglia, but not Cx3cr1⁺/Sall1^– infiltrating or vasculature-associated myeloid cells (Kim J.-S. et al., 2021). This enabled differential analysis of microglial and BAMs translatomes through a Cre-mediated RiboTag strategy (Kim J.-S. et al., 2021), highlighting strong potential of binary Cre systems for future use in high-specificity studies of microglial biology.

As a result of advances in gene editing techniques, methods of microglial depletion—based on KO animal models of essential genes in microglial survival and function—have provided a more comprehensive understanding into the roles of microglia in development, homeostasis, and pathology. Meanwhile, selective KO models of microglial genes across various contexts have helped us better understand the roles of specific genes and microglial states dependent on those genes.

Microglia, as well as monocytes and other macrophage populations, highly express the receptor for colony-stimulating factor-1 (CSF1R). CSF1R signaling through its ligands, CSF-1, and IL-34 (Wang et al., 2012), is essential for microglial development, survival, proliferation and release of proinflammatory factors (Mitrasinovic et al., 2005; Ginhoux et al., 2010; Erblich et al., 2011; Elmore et al., 2014; Green et al., 2020). As a result, CSF1R is a well-established target for microglial depletion. The loss of CSF1R in Csf1r^–/– mice leads to nearly a complete lack of microglia, with > 99% depletion at E16 and postnatal day 1 (Ginhoux et al., 2010; Erblich et al., 2011). However, Csf1r^–/– mice also show deficits in brain size and function, as well as some deficiencies of the olfactory bulb development, myelination and bone structure, leading to a premature death (Erblich et al., 2011; Green et al., 2020). It thus requires the use of inducible Cre/lox systems or pharmacological inhibitors of CSF1R (discussed below) to selectively deplete microglia in adult mice. Pons et al. (2021) used the Cre/loxP system to delete CSF1R in 3-month old mice, achieving a KO efficiency of 89%, and showed that loss of Csf1r in adulthood did not impact microglial survival. Intriguingly, deletion of Csf1r in adult APP-PS1 mice significantly ameliorated AD amyloid-beta pathology (Pons et al., 2021), highlighting the potential of this tool to elucidate the complex roles of microglia in neurodegenerative disease. Deletion of Csf1r regulatory elements, such as the fms-intronic regulatory element (FIRE), can also deplete microglia by selectively abolishing Csf1r expression in a cell- and tissue-specific manner (Rojo et al., 2019). Mice with CRISPR-mediated homozygous FIRE deletions completely lack microglia by IBA1 and P2RY12 immunostaining, as well as several other tissue-resident macrophage populations, but otherwise develop into healthy fertile adults without the systemic defects observed in Csf1r^–/– mice. Notably, flow cytometry identified retention of CD11b⁺/CD45^hi brain macrophages, which also co-expressed the perivascular macrophage marker CD169 by immunostaining (Rojo et al., 2019), suggesting a high degree of depletion specificity to microglia. Researchers have since crossed FIRE KO mice with the 5XFAD mouse model to study the effects of selectively lacking microglia in AD pathology, showing that the absence of microglia ameliorated amyloid-beta pathology but worsened cerebral amyloid angiopathy, which could be prevented by intra-hippocampal transplantation of wild-type microglial cells (Kiani Shabestari et al., 2022). The use of highly selective microglial depletion models holds great promise to elucidate microglial roles, and associated therapeutic targets, across diverse pathologies.

As previously discussed, the CX3CR1 receptor is a powerful tool to investigate neuron-microglia communication (Wolf et al., 2013), and Cx3cr1 KO models have been critical for our understanding of the fractalkine signaling pathway. Pagani et al. (2015) demonstrated that microglia-specific Cx3cr1 KO leads to acute deficits in the developing hippocampus, including impairments in the response to extrinsic ATP and altered electrophysiological properties. In mouse models of chronic stress, Cx3cr1 deficiency attenuated stress-induced alterations to microglial morphology and reduced phagocytosis (Milior et al., 2016), highlighting a key role for microglia-neuron communication in the response to stress. An additional method microglial depletion is achieved by placing the diphtheria toxin receptor (DTR) under the control of a microglial promoter using inducible Cre-loxP models, allowing for an efficient depletion when diphtheria toxin is administered. Using Cx3cr1^CreER:R26^iDTR mice, Parkhurst et al. (2013) selectively depleted Cx3cr1-expressing microglia to show a critical role for these cells in brain-derived neurotrophic factor-dependent learning and synaptic plasticity. A recent study used similar tools to assess the effects of depleting Cx3cr1⁺ microglia on kainic acid-induced status epilepticus (Wu et al., 2020). Depletion of microglia through inducible expression of diphtheria toxin, DTR or KO of the essential Csf1r in Cx3cr1⁺ cells, led to exacerbated disease pathology, increased mortality and enhanced neurodegeneration (Wu et al., 2020). However, as Cx3cr1 is also expressed on infiltrating and border-associated myeloid cells, future studies would benefit from employing more selective depletion tools such as binary Cre-based systems targeting multiple co-expressed microglial genes (Kim J.-S. et al., 2021).

Triggering receptor expressed on myeloid cells-2 (TREM2) is essential for microglial synaptic pruning and formation of normal neural circuitry during brain development by mediating phosphatidylserine-dependent phagocytosis of apoptotic neurons (Filipello et al., 2018; Scott-Hewitt et al., 2020). It is also involved in the microglial response to injury and in the recognition of soluble factors (Quan et al., 2009; McQuade et al., 2020). Microglial TREM2 is heavily implicated in disease and is required for the transformation into DAM, MgND and other pathology-associated states (Keren-Shaul et al., 2017; Krasemann et al., 2017; Deczkowska et al., 2018). TREM2 is selectively expressed on microglia within the CNS, but is found on other myeloid-derived cells such as macrophages/monocytes, dendritic cells and granulocytes in the periphery (Yaghmoor et al., 2014; Gratuze et al., 2018). Soluble forms of TREM2 increase microglial survival and proliferation and induce morphological changes during inflammatory responses (Zhong et al., 2017). Furthermore, TREM2 variants in humans have been strongly implicated in AD risk (Guerreiro et al., 2013; Jonsson et al., 2013; Sims et al., 2017). Genetic Trem2 deletion models have greatly contributed to our knowledge of its normal functions and implications in microglial biology, generally implicating TREM2 as a critical mediator of phagocytosis in development and pathological states. Trem2^–/– mice show decreased microglial numbers in the CA1 region of the hippocampus with reductions in reactive microglial morphologies, increased synaptic density and overactive excitatory neurotransmission, highlighting the importance of Trem2 in the removal of extraneous synapses during neural development (Filipello et al., 2018). These deficits are recapitulated in human iPSC-derived microglia in vitro, which following CRISPR-mediated deletion of TREM2, are unable to phagocytose apolipoprotein E or fibrillar amyloid-beta, and present impaired survival in response to stress (McQuade et al., 2020). In line with this, 5XFAD mice deficient in Trem2, or its downstream mediator Syk, display an impaired amyloid-beta plaque clearance by microglia (Wang et al., 2015, 2022), and do not develop DAM (Keren-Shaul et al., 2017; Wang et al., 2022). Similarly, in a humanized mouse model of Tau pathology, crossed mice lacking Trem2 presented augmented Tau aggregation and phosphorylation with altered morphology in IBA1⁺ microglia by IHC (Bemiller et al., 2017). TREM2 is also expressed on plaque-associated dark microglia by EM in the APP-PS1 model of AD pathology (Bisht et al., 2016), suggesting a role for dark microglia in amyloid-beta phagocytosis. However, selective deletion of Trem2 in microglia leads to an altered transcriptional profile and phenotype, temporarily improving disease in mouse models of traumatic spinal cord injury and experimental autoimmune encephalomyelitis (Gao et al., 2023). Overall, Trem2 KO models have enabled the selective depletion of pathologically relevant microglial populations, shedding light on the diverse roles of microglia in disease.

In a similar fashion, C-type lectin domain family member 7a (CLEC7a), also known as DECTIN-1, is expressed on cells of the monocyte lineage and functions in the recognition and phagocytosis of fungal and bacterial pathogens (Brown et al., 2002). CLEC7a was also shown to play a role in cancer progression through its influences on tumor-associated macrophages (Daley et al., 2017). CLEC7a is expressed in microglia, where it promotes phagocytosis upon ligation (Shah et al., 2008). CLEC7a signaling directly activates SYK-mediated pathways downstream of TREM2, enhancing phagocytosis and representing a potential therapeutic target for individuals with deficits in TREM2 function (Wang et al., 2022). In mouse models of AD pathology, brain Clec7a is significantly upregulated in DAM and MgND states (Keren-Shaul et al., 2017; Krasemann et al., 2017; Deczkowska et al., 2018). CLEC7a is also upregulated by microglia in mouse models of experimental autoimmune encephalomyelitis (Krasemann et al., 2017; Deerhake et al., 2021), with Clec7a^–/– mice displaying more severe disease, although these effects were more likely mediated by CNS-infiltrating myeloid cells than microglia (Deerhake et al., 2021). These examples highlight the potential for CLEC7a KO models to continue to enhance our understanding of microglial roles in disease conditions.

The pharmacological depletion of microglia in the CNS has been a valuable and accessible tool in order to define their role across contexts of health and disease, with the advantage of being less challenging than generating and utilizing a KO model (Barnett et al., 2021). This has been made possible due to a vast array of compounds that demonstrate the capacity to modulate microglial numbers and control their activities, notably in mediating CNS inflammation (Barnett et al., 2021).

As discussed, the use of CSF1R inhibitors is an emerging technique with demonstrated efficacy in the ablation of microglia in the CNS (Elmore et al., 2014; Fujiwara et al., 2021). The CSF1R inhibitors, such as PLX3397 and PLX5622, are CNS-permeable and selectively inhibit tyrosine kinase receptors on macrophages and microglia, although the specificity of these inhibitors varies depending on the specific compound used (Green et al., 2020; Hupp and Iliev, 2020), with PLX5622 recently shown to exert effects on peripheral bone marrow and tissue-resident macrophages (Lei et al., 2020). CSF1R inhibitors are orally bioavailable, but their routes of administration can also include intravenous or subcutaneous injection (Hupp and Iliev, 2020). Moreover, CSF1R inhibitors have been shown to reduce microglial proliferation, depleting the majority of the brain population (Elmore et al., 2014; Acharya et al., 2016; Fu et al., 2020). It has been reported that most of IBA1⁺ microglia are eliminated within the first week of administration—as assessed with staining against caspase-3, a marker for cell apoptosis. Approximately 90% of microglia can be eliminated following 7 days of PLX3397 treatment, shown using flow cytometry in whole brains of CX3CR1-GFP^+/– adult mice (Elmore et al., 2014). Effects remain with sustained administration of CSF1R inhibitors—for instance, after 21 days of PLX3397 treatment (Elmore et al., 2014). However, microglial repopulation is notable between drug withdrawal and the 3-day recovery time point. This highlights the critical period of 48–72 h for microglial repopulation after CSF1R inhibition (Elmore et al., 2014; Henry et al., 2020; Shi et al., 2022). PLX3397 has undergone exploration in humans; however, its efficacy has not yet achieved that observed in mice. In a phase II clinical trial for recurrent glioblastoma, a daily oral dose of 1,000 mg of PLX3397 was well-tolerated and successfully crossed the blood-brain barrier of 37 patients (mean age of 58.5; 25 males and 12 females). Nevertheless, it demonstrated no effectiveness in reducing tumor size or microglial density within the tumor tissue. IHC for IBA1 was conducted and used to quantify cell numbers and staining intensity. No statistically significant differences were observed in IBA1⁺ microglia between pre-study (archival) and on-study (surgical) samples, whether treated with PLX3397, following the standard of care with radiation therapy complemented with temozolomide chemotherapy, or when compared to historical control samples (Butowski et al., 2016).

Interestingly, the elimination of microglia using CSF1R inhibitors was reported to prevent amyloid-beta plaque formation and disease progression in the 5XFAD and APP-PS1 mouse models of AD pathology (Olmos-Alonso et al., 2016; Spangenberg et al., 2019), and to reduce accumulation of pathogenic Tau in the Tg2541 tauopathy model, extending survival of female but not male mice (Johnson et al., 2023). Likely microglia, BAMs have the fully capacity to repopulate via local self-renewal after depletion via PLX3397 (Van Hove et al., 2019). These studies highlight CSF1R as a potential target to deplete microglia in therapeutic approaches for neurodegenerative diseases.

Minocycline, a drug initially developed to treat bacterial infections, has been shown to control and deplete inflammatory cells in the CNS (Garrido-Mesa et al., 2013; Scholz et al., 2015). Recently, it was reported that tetracyclines not only display anti-microbial activity, but also act as anti-inflammatory and anti-apoptotic agents while impairing proteolysis and angiogenesis (Bernier and Dréno, 2001; Scholz et al., 2015). Moreover, minocycline has emerged as a potential neuroprotective agent in experimental models for several CNS disorders, such as brain ischemia (Yrjänheikki et al., 1998), brain injury (Sanchez Mejia et al., 2001), Parkinson’s disease (Du et al., 2001; Thomas and Le, 2004), Huntington disease (Chen et al., 2000) and AD (Choi et al., 2007). Notably, minocycline appears to exert an anti-inflammatory influence on microglia (Bassett et al., 2021; Long et al., 2023). To exemplify, in a mouse model of subarachnoid hemorrhage (SAH), minocycline treatment was administered daily until day 7 post-SAH and continued every second day until day 14. IHC for IBA1 and a functional phagocytosis assay, employing glucose-coated beads cross-linked with a lipid bilayer and co-localized with IBA1⁺ cells in brain slices, indicated a significant reduction in the phagocytic activity of microglia/macrophages. This reduction was associated with decreased spatial interaction with neurons (NeuN⁺) and a decline in neuronal apoptosis (TUNEL⁺/NeuN⁺) (Blecharz-Lang et al., 2022).

The efficiency of minocycline in depleting microglia can vary depending on the specific conditions of the study, such as the dosage, route of administration and duration of treatment, however, overall it has been shown to be a relatively effective method to modulate microglial populations, typically achieving 50–90% depletion (Blecharz-Lang et al., 2022). However, minocycline also exhibits broad off-target effects in the CNS, and thus results from minocycline-based depletion studies should be considered with caution (Strahan et al., 2017).

Likewise, clodronate-containing liposomes (Clod-Lips) constitute an effective approach to deplete macrophage-like cells across tissues (Rooijen and Sanders, 1994; Moreno, 2018). Initially discovered as an osteoclast inhibitor, clodronate, following phagocytosis and entry into the macrophage cytosol, perturbs iron metabolism and inhibits mitochondrial respiration through its actions as an ATP analog, leading to apoptosis (Lehenkari et al., 2002; Moreno, 2018; Opperman et al., 2019). The efficiency of clodronate depends on its route of administration; intrarectal and intraperitoneal administration deplete around 50% of the macrophage population, while 90% can be achieved through the intravenous route (Weisser et al., 2012). A major drawback to Clod-Lips is their requirement for direct infusion into the CNS, as well as their demonstrated off-target effects on blood vessels (Han et al., 2019). Similarly, Takagi et al. (2020) demonstrated that Clod-Lips markedly attenuated the activation of mouse brain circuits for TLR2-mediated inflammation in a hypothermia mouse model.

Mac-1-saporin (Mac-1-sap), a chemical conjugate of a CD11b monoclonal antibody and the ribosome-inactivating protein saporin, is another pharmacological option for microglial ablation that has been shown to deplete CD11b⁺ cells in mouse models (Acharjee et al., 2018). Although not specific to microglia, targeting all CD11b⁺ cells, intrathecal injection of Mac-1-sap depletes around 50% microglia in the CNS (Yao et al., 2016). Following depletion, repopulating microglia were shown to respond to injury, re-establishing normal function after ablation (Yao et al., 2016). In vitro work further investigated the depletion of microglia from mouse hippocampal cultures in a model of ischemia-like oxygen-glucose deprivation. In this context, seven-day treatment with Mac-1-sap drove neural apoptosis and astrogliosis one day after oxygen-glucose deprivation (Montero et al., 2009), supporting the importance of microglia in the response to ischemic injury.

Several important microglial fate mapping tools were developed specifically aiming to decipher the origin of microglia. In these approaches, early embryonic progenitor cells are labeled based on specific gene expression patterns during development, and their progeny is followed to identify the progenitor cells giving rise to microglia and the key transcription factors regulating these processes. In a landmark study, Ginhoux et al. (2010) used fate-mapping techniques to determine that microglia develop from yolk sac-derived myeloid progenitors that arise prior to E8 in mice. In this study, Cre was co-expressed with Runx1, a transcription factor expressed in early hematopoietic progenitors, in mice harboring a Rosa26-fl-STOP-fl-YFP construct (Runx1^MerCreMer:R26^YFP). Tamoxifen administration at E7.5 labeled all subsequent microglia as YFP⁺, demonstrating that microglia develop from yolk sac-derived Runx1-dependent progenitors (Ginhoux et al., 2010). This finding was later confirmed using a similar approach (Hoeffel et al., 2015), and Runx1-based fate mapping was again employed to show that retinal microglia share a common yolk-sac origin with brain microglia (O’Koren et al., 2019). Additional research further described microglial development by co-expressing CreER with CSF1R, a constitutively expressed receptor in microglia, monocytes, and macrophages, in R26^YFP mice. Using tamoxifen administration to induce YFP labeling at different time points in development further revealed that microglia develop independently from bone-marrow hematopoietic stem cells (Schulz et al., 2012; Gomez Perdiguero et al., 2015; Hoeffel et al., 2015). Another study took advantage of the transcription factor Kit, expressed in early yolk sac progenitors, using Kit^MerCreMer:R26^YFP mice to show that only microglia, along with a subset of Langerhans cells, develop from early (E7.5) Kit⁺ erythromyeloid progenitors in the yolk sac (Sheng et al., 2015). Thus, microglial fate mapping tools have been crucial to building our current knowledge of microglia as an ontogenetically distinct macrophage lineage.

Several key questions in modern microglial biology relate to dissecting the roles of microglia from other CNS myeloid cells and distinguishing among heterogeneous states within the microglia. Fate mapping studies have been critical in addressing these fundamental concepts. O’Koren et al. employed a combination of fate mapping and 12-color flow cytometry to definitively distinguish between retinal microglia and monocyte-derived macrophages, which both express CX3CR1, in CX3CR1^YFP–CreER:R26^RFP mice. Following tamoxifen administration, both lineages expressed RFP, however, short-lived monocyte-derived macrophages were replaced by RFP^– cells over a subsequent “washout” period, while long-lived microglia retained expression of RFP (O’Koren et al., 2016). Using a similar mouse model, the time course of microglial and macrophage infiltration into the various compartments of the eye was recently mapped at high spatiotemporal resolution (Wieghofer et al., 2021). More recently, researchers identified prominin 1 as a marker of committed myeloid progenitors in the adult CNS under homeostatic conditions, and employed a novel fate mapping approach using Prom1^CreERT2:R26^tdT mice to demonstrate that Prominin-1 (Prom1)⁺ progenitors also contribute to microglial proliferation during homeostasis (Prater et al., 2021).

Heterogeneity within the microglial population is an emerging concept that, until recently, could not be explored with available fate-mapping tools. To address this issue, Tay et al. (2017) created the “Microfetti” mouse strain, a cross between the existing CX3CR1^CreER line and a R26^Confetti strain, which contains a four-color reporter system. In Microfetti mice, tamoxifen administration causes random expression of one of four fluorescent reporters in each cell thus creating up to ten distinct color combinations in the homozygous state (Tay et al., 2017). Using mathematical models to evaluate the likelihood of spatial association of same-color cells due to chance versus cell division, the authors achieved detailed tracing of microglial cells and their progeny at single-cell resolution. This revolutionary approach revealed that microglial proliferation is a stochastic process governed by regional environmental cues during homeostasis (Tay et al., 2017).

Microglia, as the innate immune cells of the CNS, play diverse and complex roles in the response to tissue injury, infection and in chronic pathological states. Fate mapping studies have allowed researchers to discern the distinct roles of microglia, BAM subsets, and infiltrating peripheral myeloid cells. For example, the source of repopulating microglia following experimental CSF1R-mediated depletion has been extensively debated. Huang et al. (2018) employed a fate mapping approach to address this issue. Using CX3CR1^CreERT2:R26^tdT mice given tamoxifen early in postnatal life, they showed that all repopulating cells were tdT⁺ and were thus exclusively derived from surviving microglia, rather than peripheral monocytes (Huang et al., 2018). Subsequent research using a similar fate mapping model has traced these repopulating microglia to a MAC2⁺ progenitor population with an immature gene expression signature that displayed resistance to CSF1R inhibition (Zhan et al., 2020). Other groups have employed comparable techniques to replicate these results in the retina (Zhang et al., 2018; O’Koren et al., 2019).

Single-cell-resolution fate mapping, using the CX3CR1^CreER:R26^Confetti model described above, showed that microglia proliferate in waves of clonal expansion under pathological conditions, and that homeostasis is restored through both migration away from the site of pathology and apoptosis (Tay et al., 2017). A recent study mapped the dynamics of microglial populations in an inducible mouse model of AD pathology (Wu et al., 2021). Using a Kit^MerCreMer:R26^YFP system with tamoxifen-mediated YFP induction periodically over 8 months, microglia, BAMs and parenchymal macrophages were shown to remain relatively stable throughout disease progression, while other myeloid cells were rapidly and continuously replaced by bone marrow-derived monocytes (Wu et al., 2021). In cuprizone-induced demyelination, fate mapping using CX3CR1^{CreER–iresGFP}:R26^dsRed mice also revealed that microglia, not bone marrow-derived macrophages, are the primary cells involved in the remyelination response (Marzan et al., 2021).

Fate mapping in mouse models of retinal pathology have also provided intriguing findings. In a CX3CR1^CreERT2:R26^tdT mouse model of retinal ischemia-reperfusion, proliferation of resident microglia and recruitment of peripheral monocytes were found to play similarly important roles in the inflammatory response (Ahmed et al., 2017), while a study in a similar model of optic nerve crush injury found that microglia, not peripheral monocytes, were the primary cells recruited to the site of injury (Heuss et al., 2018).

As diverse microglial fate mapping systems have proven useful for the analysis of microglial dynamics in development, homeostasis and disease, data has become available on the efficacies of various mouse strains in targeting microglia. Following the advent of Cx3cr1^Cre-based models, their specificity for microglia was assessed and directly compared to earlier Lyz2^Cre and Cd11c^Cre mice, through direct crosses to R26^YFP strains (Goldmann et al., 2013). While Cx3cr1^Cre-targeted microglia with higher sensitivity and specificity than Lyz2^Cre or Cd11c^Cre, recombination was also observed in CX3CR1⁺ infiltrating peripheral myeloid cells. However, long-lived microglia, which retained a stable expression of YFP, could still be distinguished after a washout period of several weeks, while short-lived peripheral CX3CR1⁺ monocytes turned over rapidly and were replaced by YFP^– cells (Goldmann et al., 2013). Similar studies validated the utility of this approach as a means of distinguishing CX3CR1⁺ infiltrating monocytes and dendritic cells from resident parenchymal microglia (O’Koren et al., 2016; Mrdjen et al., 2018; Jordão et al., 2019), although it has been noted that some BAM subsets, such as those residing in the perivascular spaces meninges and CP, are replaced more slowly than other subsets and thus cannot be readily distinguished through this approach alone (Mrdjen et al., 2018; Jordão et al., 2019). Later work also established that many BAM subsets, in fact, express CX3CR1 (Goldmann et al., 2016). Accordingly, off-target recombination in non-parenchymal CNS macrophages is observed in CX3CR^CreER systems (Chappell-Maor et al., 2020; Faust et al., 2023).

Recent advances in single-cell transcriptomics have identified a battery of novel microglia-enriched genes, several of which have been leveraged to generate Cre-based tools with applications to fate mapping. Examples include Sall1 (Buttgereit et al., 2016), Tmem119 (Kaiser and Feng, 2019), P2ry12 (McKinsey et al., 2020), and Hexb (Masuda et al., 2020). Sall1^CreER targets microglia with high sensitivity and specificity, notably showing no recombination in CNS-infiltrating peripheral myeloid cells and nearly all non-parenchymal BAMs with the exception of a small subset of CP BAMs (Buttgereit et al., 2016; Van Hove et al., 2019; Chappell-Maor et al., 2020). However, Sall1 is a critical regulator of microglial homeostasis (Buttgereit et al., 2016), likely rendering this approach less suitable for microglial fate mapping in pathological states. Furthermore, off-target recombination was recently reported in neuroectoderm-derived CNS cells in Sall1^CreER mice (Chappell-Maor et al., 2020). Likewise, the Tmem119^CreERT2 line (Kaiser and Feng, 2019) labeled adult microglia with high specificity, although some recombination was noted in leptomeningeal macrophages and other non-IBA1⁺ cells thought to be perivascular or meningeal fibroblasts (Kaiser and Feng, 2019; McKinsey et al., 2020). The P2ry12^CreER mouse line was developed with the goal of preserving endogenous expression of the functional P2RY12 receptor, which plays important roles in nucleotide-sensing during the microglial response to injury, seen as a benefit in contrast to the reduced expression of Sall1 in Sall1^CreER knock-in strains (McKinsey et al., 2020). The P2ry12^CreER specifically and efficiently labeled embryonic as well as adult microglia, offering an advantage over the Tmem119^CreERT2 construct, which mediated off-target recombination in CD31⁺ endothelial cells during development (Kaiser and Feng, 2019; McKinsey et al., 2020). No off-target recombination was observed in perivascular macrophages or in circulating blood cells, although a subset of CP macrophages were labeled, possibly corresponding to the microglia-like Kolmer’s epiplexus BAMs described by Van Hove et al. (2019) and McKinsey et al. (2020). However, like Sall1, Tmem119, and P2RY12 are strongly associated with homeostatic microglial signatures (Kaiser and Feng, 2019; McKinsey et al., 2020). In fact, it was proposed that Tmem119-based systems could potentially be studied as sensors of disease states, as the expression of this gene may be perturbed under pathological conditions (Ruan et al., 2020). Hexb-based genetic tools have also recently been developed (Masuda et al., 2020). Massively parallel single-cell RNA-sequencing identified Hexb as a highly enriched gene in microglia that, unlike Tmem119, Sall1, or P2ry12, is consistently expressed across conditions of homeostasis and disease. Hexb^CreERT2:R26^YFP mice showed highly specific recombination in microglia, with distinct advantages over Cx3cr1- and Sall1-based systems (Masuda et al., 2020). However, although fate mapping validation experiments showed that Hexb^CreERT2 did not label other CNS macrophage subsets or infiltrating peripheral monocytes, a measurable subset of microglia were not labeled (Masuda et al., 2020). Additionally, a recent study demonstrated that Tmem119^CreER and Hexb^CreER lines show low recombination efficiency relative to the established Cx3cr1^CreER based tools (Faust et al., 2023). Recombination efficiency was shown to depend on the genetic distance separating the two loxP sites, with shorter inter-loxP distances promoting efficient recombination (Faust et al., 2023). Thus, further optimization of Tmem119-, Sall1-, Hexb-, P2ry12-, and Cx3cr1-based tools using evidence-driven methods will likely be required to exploit their full potential.

The previously described CX3CR1^CreER:R26^Confetti model offers a unique advantage in that fate mapping can be achieved with single-cell resolution (Tay et al., 2017). This has allowed detailed analyses of microglial dynamics under homeostatic and pathological conditions, greatly enhancing our ability to examine heterogeneity within the microglial population and track the migration and proliferation of individual cells (Tay et al., 2017). In the context of a recent explosion of microglial single-cell transcriptomics studies, approaches of this nature hold great promise as key techniques in the study of microglia at an unprecedented resolution.