Microglial cells are the innate immune cells of the brain and key players in maintaining the homeostasis of the central nervous system (CNS) (Sierra et al., 2016). Microglia originate from erythro-myeloid precursors in the yolk sac and migrate to the brain around embryonic day 9.5 in mouse (Ginhoux et al., 2010), while they colonize the human cerebrum between the 4th and 24th week of gestation (Menassa and Gomez-Nicola, 2018). Their ontogeny, together with their slow turnover, which differentiate them from most other hematopoietic lineages in adult individuals (Réu et al., 2017), as well as the local environment in the CNS, make microglia a distinct immune cell population (Sousa et al., 2017). Until approximately 20 years ago, microglia have been considered as a resident resting cell type of the healthy CNS able to react to pathogens or toxic elements. However, this paradigm has shifted into the concept of “surveillant” and “supporting” microglia exerting additional multiple functional roles, such as neuromodulation and phagocytosis (Gomez-Nicola and Perry, 2015). For example, during development microglia contribute to building the neuronal circuit through synaptic pruning and stripping, phagocytosis of dying neurons and secretion of neurotrophic factors (Paolicelli et al., 2011; Ekdahl, 2012; Schafer et al., 2012; Paolicelli and Ferretti, 2017; Scott-Hewitt et al., 2020). Further, it has recently been shown that an ATP-dependent microglia-driven negative feedback mechanism operates similarly to inhibitory neurons and is essential for protecting the brain from an excessive activation (Badimon et al., 2020). Taken together, due to their multiple critical functional roles in the homeostatic brain, various neurological disorders, including neurodegenerative diseases and brain tumors, implicate microglia. Briefly, in Parkinson’s disease (PD) and Alzheimer’s disease (AD), a mix of beneficial and detrimental roles of microglia have been suggested (Wyss-Coray and Mucke, 2002; Glass et al., 2010; Bodea et al., 2014; Joers et al., 2017; Salter and Stevens, 2017; Wolf et al., 2017; Duffy et al., 2018). For example, in AD microglial cells have been associated with the phagocytosis and degradation of amyloid-ß plaques, but the subsequent excessive release of cytokines is supposed to contribute to neuronal loss (Salter and Stevens, 2017). Similarly, in PD, where microglia are able to recognize and engulf alpha-synuclein, but the concomitant release of reactive oxygen species (ROS) or pro-inflammatory mediators can actively contribute to neurodegeneration (Glass et al., 2010). In brain tumors, microglia, along with tumor-infiltrating macrophages, constitute the predominant immunological cell types (Graeber et al., 2002) and have been shown to affect tumor progression as well as patient survival (Morimura et al., 1990; Gieryng et al., 2017; Sorensen et al., 2018). Indeed, tumor-associated microglia/macrophages (TAMs) are key players along tumor development by contributing to the establishment of a tumor-supporting microenvironment (Hambardzumyan et al., 2016; Grabowski et al., 2020; Maas et al., 2020).

In this context, the hypothesis that several microglial cell subsets exist in the brain has gain momentum in the recent years and microglial heterogeneity has been addressed from different points of views, including morphology, cellular density, proliferation capacity as well as transcriptional and proteomic signatures (De Biase and Bonci, 2018; Silvin and Ginhoux, 2018). Additionally, significant advances have been made by taking advantage of the recently developed single-cell technologies, including RNA-sequencing and mass cytometry (CyTOF). Indeed, several studies using these approaches have now confirmed that microglial cells represent a complex population constituted by different subsets, both in the healthy and diseased brain, displaying specific neuroimmunological adaptations in a context-dependent manner (Stratoulias et al., 2019; Masuda et al., 2020; Provenzano et al., 2020).

Here, we will review the main breakthrough studies that helped to elucidate microglia heterogeneity in the healthy and diseased brain. We suggest that unraveling the cellular and molecular mechanisms underlying specific microglia subpopulations might contribute to uncover new therapeutic targets for brain disorders with an immunological component, including neurodegenerative diseases and tumors.

An extensive characterization of microglial heterogeneity encompassing fundamental aspects, including development, gender specificities and spatial distribution has been conducted in the healthy brain. Hence, along this chapter, we chronologically review the main studies that contributed to acquire the current knowledge of microglia diversity under homeostatic conditions (Figures 1, 2).

Lawson et al. (1990) have conducted the first study addressing microglial heterogeneity in the 90’s. The authors showed that, in the mouse brain, microglial ramifications and cell shapes were region-dependent. By using the macrophage marker F4/80, they also reported divergent microglial densities across specific brain regions, with higher density in the hippocampus, thalamus and amygdala compared to the cortex and cerebellum (Lawson et al., 1990). Concomitantly to this study in mice, Mittelbronn et al. (2001) studied microglial density across human brain regions using myeloid-specific-immunological markers, including CD68, MHC-II, and IBA1. In this study, the authors described a higher content of microglial cells within the white-matter when compared to gray-matter (Mittelbronn et al., 2001). Seven years later, de Haas et al. (2008) took advantage of ex vivo flow cytometry analyses to investigate regional differences in the expression of immunoregulatory proteins across the mouse spinal cord and various brain regions, such as the cerebral cortex, hippocampus, cerebellum or striatum. Interestingly, the authors described that CD40 was overexpressed by microglia located in the cerebellum when compared to cortical microglia. Similarly, CD45 or CXCR3 were described to be less expressed in the hippocampus in comparison to the other regions, while TREM2 was differentially expressed between microglia in the cerebellum and cortex (de Haas et al., 2008). Later on, Doorn et al. (2015) investigated baseline differences in microglial expression of genes in brain regions associated with PD, including substantia nigra, striatum, olfactory bulb, hippocampus, or amygdala of rats. The authors did not detect differences in the expression levels of Aif1, Cd11b, or Tlr2 genes in microglia isolated from those regions. However, the expression levels of the phagocytic and pro-inflammatory markers Cd68 and Il1b were higher in the olfactory bulb compared to the other brain regions. Besides, Cd68 expression was higher in the striatum compared to the amygdala and Tnf was overexpressed in the substantia nigra (Doorn et al., 2015).

Along with a deeper understanding of microglial heterogeneity across brain regions, the raise of single-cell transcriptomic technologies has been a breakthrough toward further revealing the cellular diversity of the brain at single-cell resolution. In this context, a pioneer work from Tasic et al. (2016) enabled to build a cell taxonomy atlas of the murine cortex identifying up to 49 transcriptomic cell types, 7 of them being non-neuronal cell types, including microglia. In the same year, two key studies helped to understand microglia diversity across brain regions, although not at single-cell resolution yet. More specifically, Grabert et al. (2016) conducted a large RNA-sequencing analysis of isolated microglial cells from the mouse cortex, hippocampus, striatum, and cerebellum. This study demonstrated that the most variable gene ontology terms discriminating the analyzed brain regions were related to metabolism and immune regulation. Besides, microglial transcriptomic heterogeneity clustered into three different signatures: the “cortex, hippocampus, and striatum,” the “cerebellum and hippocampus,” and the “cerebellum” transcriptomes. Importantly, the authors showed that microglia from the cerebellum exhibit a specific “immune-vigilant” transcriptional signature when compared to the others brain regions (Grabert et al., 2016). In addition, Verdonk et al. (2016) expanded the knowledge on regional heterogeneity by shedding light on microglia morphology using an automated method based on 3D reconstruction. With this technique, microglial morphologies have been analyzed according to the complexity of primary ramifications (CI, complexity index) and the total 2D area covered by ramifications (CEA, covered environment area). In line with the transcriptomic findings obtained by Grabert et al. (2016), microglial morphology from the cerebellum was the most diverse. Indeed, cerebellar microglia had smaller CI and CEA in comparison to microglia from the hippocampus, frontal cortex and striatum, which exhibit similar cell body and cellular area (Verdonk et al., 2016). Still in 2016, the first single cell study addressing microglia diversity along development was published by Matcovitch-Natan et al. (2016) who specifically defined early-stage cycling (e.g., Dab2, Mcm5, Lyz2), synaptic pruning (e.g., Crybb1, Csf1, Cxcr2), and adult immune surveillant (e.g., MafB, Cd14, Mef2a) microglia.

It was in 2017 when De Biase et al. (2017) investigated microglia diversity across different areas of the basal ganglia, including the nucleus accumbens, ventral tegmental area, substantia nigra pars compacta, and substantia nigra pars reticulata, using a combination of morphological and transcriptomic analyses. The authors observed that microglia in the striatum and the nucleus accumbens displayed a higher complexity when compared to the others sub-regions. Further, they reported variability in terms of microglial density within the basal ganglia, detecting a higher number of microglial cells within the substantia nigra pars reticulata when compared with the substantia nigra pars compacta and the ventral tegmental area. Interestingly, they also reported a uniform density of all the other cell types within the basal ganglia, interrogating the origin of the region-specific microglial proliferation capacities. Lastly, they described different microglial transcriptional signatures, especially within the ventral tegmental area, where genes related to mitochondrial function, metabolism, oxidative signaling, or lysosomal function were differentially expressed when compared to the other regions (De Biase et al., 2017).

In 2018, two major studies addressing sexual dichotomy at the transcriptional level have been conducted. In both cases, biological processes related to the immune phenotype (e.g., cytokine production, expression of antigen-presenting cell markers and purinergic receptors) were over-represented in male microglia (Guneykaya et al., 2018; Villa et al., 2018). Further, female microglia have been linked to neuronal processes (e.g., perpetuation of neuronal transmission or promotion of neuroprotective mechanisms), as well as have been described to be more susceptible to microbiota modifications (Villa et al., 2019).

Doubtless, 2019 has been the year when microglial heterogeneity has been mostly elucidated at single-cell resolution. For example, Bottcher et al. (2019), by using mass-cytometry, described nine different human microglial substates across the subventricular zone, thalamus, cerebellum, and the temporal and lateral lobes, with the subventricular zone displaying a more distinct signature compared with the other brain regions. In mice, Hammond et al. (2019), by using single cell technology, found high microglial heterogeneity in young mice (E14.5 and P5), with eight different substates of microglia, for which the expression of markers such as Arg1, Rrm2, Hmox1, or Spp1 differed. In addition, the authors aimed to identify gender-specific differences of microglia across three different ages (E14.5, P4/P5, and P100), but did not detect main differences between males and females, except for the expression of chromosome-specific genes, such as Eif2s3y and Xist (Hammond et al., 2019). In addition to the study of Hammond et al. (2019), Li et al. (2019) described the so-called proliferative-region-associated microglia (PAM), a specific microglial subtype located in the developing cerebellar white matter and corpus callosum, which is characterized by the expression of Spp1 and Gpnmb. Lastly, Masuda et al. (2019) conducted a more detailed single-cell study, mainly focused on microglia, in which their heterogeneity has been addressed across specific brain regions, such as the corpus callosum, cerebellum, cortex, hippocampus, and facial nucleus at different mouse ages. The authors identified 10 main substates during development with differences in the expression of microglial markers, such as Tmem119, Malat1, Lamp1, or Apoe (Masuda et al., 2019). Notably, in the same study, four different substates of microglia have also been described in the cortex of the human brain (Masuda et al., 2019). In parallel, microglial diversity between the gray and white matter has been studied at the transcriptomic level in the human brain (van der Poel et al., 2019). While microglia located in the gray matter were enriched in genes related to cytokine signaling, microglia in the white matter were characterized by genes involved in chemotaxis and inflammatory responses (van der Poel et al., 2019). In the same year, an automated method named MIC-MAC (Microglia and Immune Cells Morphologies Analyzer and Classifier) enabled to evaluate microglial density and morphologies at single-cell level in the hippocampi of human and mouse brains. This method allowed the clustering of microglial subpopulations based on their similarities in a 3D environment and assisted the identification of 10 different microglial subsets within the mouse and the human hippocampus, revealing a unique subset of human microglia (Salamanca et al., 2019). Still regarding microglia diversity in the human brain, Sankowski et al. (2019) recently applied high dimensional techniques and identified up to eight clusters with differential expression of microglia core genes (e.g., CX3CR1 and TMEM119), genes related to major histocompatibility complex II (e.g., HLA-DRA and CD74) as well as chemokines and cytokines (e.g., CCL2 and IL1B). In the same study, the authors explored regional-associated differences in the temporal lobe between the gray and white matter and detected higher expression levels of immune genes in microglia located in the latter (Sankowski et al., 2019). Later in 2019, Geirsdottir et al. (2019) studied microglia diversity across a various range of species across the evolutionary tree, including mouse and human. The authors used a combination of single cell transcriptomics and 3D reconstruction to study microglia morphology. In this particular study, the authors showed a conserved morphological pattern of parenchymal microglial cells across the species. However, they called attention on variations observed in terms of dendrite length, number of segments, branch points, and terminal points between the cortical and cerebellum microglia in mice and humans. From the transcriptomic point of view, microglia express a set of core genes that are conserved across species, with human microglia displaying a more pronounced heterogeneity than other mammals. Lastly, microglial genes related to the complement pathway, phagocytosis, or genes implicated in neurodegenerative diseases differ between rodents and primates (Geirsdottir et al., 2019).

Microglial cells scan the brain parenchyma and react to specific threats to avoid a disturbance of the critical, fine-tuned activities of the CNS. As mentioned previously, the phenotypic analysis of microglia in the healthy brain parenchyma revealed specific poised subsets, which might eventually support or harm the neuronal network under specific vulnerabilities. For example, this is all the more important for the understanding of CNS disorders exhibiting regional-specific and cellular pathological hallmarks, as seen in various neurodegenerative diseases, including AD (entorhinal cortex and hippocampus) and PD (nigrostriatal pathway). Therefore, in this chapter we extend the study of microglial heterogeneity in a disease-associated context. More specifically, we tackle microglia diversity associated with neuroinflammatory and neurodegenerative diseases as well as with brain tumors (Figure 3).

Undoubtedly, the ability to acquire different resting and activated phenotypes confers microglia the advantage to be a plastic and adaptive cell type in the CNS. Along with this review, we have highlighted a large number of studies demonstrating that microglia are far from being a resting or homogenous cell population. For example, in the healthy mouse brain, microglia heterogeneity across various regions has been described in terms of density, morphology, molecular signatures, and metabolism (Tan et al., 2020). For translational purposes, it will be critical to consider that microglia heterogeneity has been suggested to be even higher in the human brain compared to mouse (Prinz et al., 2019). Notably, regarding neurodegenerative diseases little is known about microglial heterogeneity in PD. Hence, future efforts will need to be directed at understanding if specific microglia subsets might differently contribute to PD pathology. In the context of brain tumors, with the difficulty to define reliable markers deciphering macrophages and microglia, the majority of the transcriptomic studies have been focusing on defining a glioma-associated microglial signature in different models, but not assessing their heterogeneity within the tumor mass (Darmanis et al., 2017; Walentynowicz et al., 2018; Maas et al., 2020). The study by Sankowski et al. (2019), being the first one highlighting specific heterogeneous signatures of microglial cells in GBM, calls attention on the lack of knowledge on microglial heterogeneity in that context. Hence, further studies would need to be directed at understanding the implication of different microglial subsets in glioma development and progression (Sankowski et al., 2019). Similarly detrimental, however, much more frequently than primary brain tumors, also brain metastases, at least experimentally, show a relevant contribution of microglia in the establishment and progression of secondary brain tumors, therefore constituting a target for future treatment strategies.

Taken together, microglial variety embraces fundamental aspects, such as spatial-temporal organization, which is present in the healthy and diseased brain. Thanks to the development of high-throughput technologies, including single-cell approaches, different microglial subsets have been unraveled, indicating that microglia are able to adapt to specific environments across particular niches in the healthy brain. In addition, single-cell analyses have been conducted to study microglia associated with inflammation and neurological disorders untangling specific subsets of cells that might differently contribute to each specific pathology (Masuda et al., 2020). Notwithstanding, a crucial aspect that would need to be tackled in future studies would be to understand the functional implication of specific microglial subpopulations across particular neurological diseases, which might enable to explore novel avenues to target neuroinflammation and microglial cells in a specialized context-dependent manner.

OUH and AM conceived the manuscript. OUH, LR, and AM wrote the manuscript. OUH created the figures. OUH, LR, MM, and AM critically revised and approved the final version of the manuscript. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.