Microglia and astrocytes are resident glial cells that influence each other under homeostatic conditions (20) but also when the CNS is affected by pathology.

In vitro, the classical inducer of neuroinflammation LPS stimulates microglia to produce a secretome enriched in NFκβ-regulated molecules including IL-1, TNF and C1q. The microglia-conditioned medium was sufficient not only to turn non-activated into activated astrocytes, assessed by the gained expression of GFAP, but also to induce the production of an unidentified secreted factor/s by astrocytes that compromised the viability of neurons and oligodendrocytes (21) (Figure 1A). This particular behavior of activated astrocytes with neurotoxic properties has been suggested to be present in patients with neurodegenerative (Alzheimer, Huntington, Parkinson, amyotrophic lateral sclerosis) and autoimmune (multiple sclerosis) CNS disorders. This hypothesis was supported by the increased levels of three proteins (C3, CBF, and MX1) that were initially found to be upregulated in a transcriptional signature of microglia-stimulated astrocytes (21).

A similar crosstalk between astrocytes and microglia was probed to be involved in some rare forms of Parkinson with mutations in the orphan receptor NURR1. Under normal circumstances NURR1 blocks the activation of NFκβ-dependent genes, a function that is lost in these patients with Parkinson disease. The combination of engineered inactivating Nurr1 mutations in glial cells with a background of increased inflammation (i.e., LPS treatment) lead to the death of dopaminergic neurons, which is a hallmark of Parkinson (22). The molecular analysis of this cross-talk probed that Nurr1-mutant astrocytes had an augmented response to microglia-derived TNFα and IL1β involving a sustained occupancy of the iNOS promoter by p65, thus secreting nitric oxide (NO) at levels that might be responsible for compromising neuronal viability (22) (Figure 1A).

The dependency of astrocyte activation on microglia behavior, was also validated in vivo in a mouse model of experimental autoimmune encephalomyelitis (EAE). Activation of the aryl hydrocarbon receptor (AhR) in microglia promotes the expression of TGFα. On the contrary, the absence of AhR signaling limits the expression of the NFκβ negative regulator Socs2, which increases the secretion of NFκβ-dependent molecules such as VEGFB (Figure 1A). Microglia secretomes enriched in either TGFα (when AhR is activated) and VEGFB (when AhR is inactivated) induced opposite transcriptomic responses when added to astrocytes by decreasing or inducing, respectively, the expression of Ccl2, Nos2 and IL1b. Some of the deregulated genes were enriched in activated astrocytes with the ability to compromise the viability of neurons and oligodendrocytes (21). In fact, when AhR was targeted in the context of EAE, disease worsened. Furthermore, targeting Ccl2, Nos2 and IL1b using cell-specific loss of function approaches either in microglia or astrocytes improved EAE outcome (23). AhR could be activated by tryptophan-derived metabolites (24). Since tryptophan is an essential amino acid provided by diet that is processed by the gut microbiome, this suggests the possibility that diet and the intestinal microbiota could have an impact on neuroinflammation. Interestingly, depleting tryptophan from the diet mimicked the phenotype of targeting AhR in microglia thus worsening EAE. Adding back the amino acid in the diet rescued the phenotype but only when the AhR receptor was present (23).

In summary, evidence exists about the critical influence of microglia on astrocytes in CNS disorders. The degree of activation of a NFκβ-dependent secretome in microglia defines the consequences on astrocytes. Microglia-activated astrocytes could worsen disease outcome by their negative influence on neuron and oligodendrocyte viability. Although the influence of microglia on astrocytes have been probed, whether astrocytes could influence microglia is less well-characterized (25).

Monocytes are excluded from the healthy brain. However, when the brain gets injured, CCR2+ circulating monocytes access the parenchyma (26, 27). As a key component of the BBB, astrocytes are one of the first cell types encountered by infiltrating peripheral immune cells, which provides the glial cell a strategic position to control this transit.

Traumatic brain injury has an impact in the viability of astrocytes located in the proximity of the damaged area. Simultaneously to the decrease in astrocytes, there is an increase in the infiltration of CCR2+ monocytes, which suggests that these cell types could influence each other. Juxtavascular astrocytes are a subpopulation that interacts physically with brain vessels and proliferation upon damage (13, 28). Although this subpopulation of astrocytes has been shown to correlate with a specific developmental origin, they were not characterized at the molecular level. Recently, juxtavascular astrocytes have been shown to preferencially activate AhR. Given that AhR blocks the production of CCL2, a strong chemokine for CCR2+ monocytes, this subpopulation of astrocytes acts as a selective barrier modulating the access of peripheral cells into the brain parenchyma (28).

Monocytes also influence astrocytes. If traumatic injury is generated in a mouse without CCR2+ monocytes, higher numbers of proliferative astrocytes are detected, suggesting a deleterious influence of infiltrated monocytes on the proliferation of juxtavascular astrocytes (28). Interestingly, in spite of the increased proliferative rates of these astrocytes, the glia scar and extracellular matrix deposition surrounding the damage was reduced and consequently, better neuronal recovery was detected (28). This finding illustrates the importance of defining at the molecular level newly established cell-to-cell interactions that occur once peripheral cells from the innate immune system infiltrate the brain. It also illustrates the importance of characterizing astrocyte heterogeneity given the impact that specific astrocyte subtypes have on disease progression (28).

In spite of the evidences presented in other brain pathologies, the crosstalk between astrocytes and macrophages had been barely explored in brain tumors. This is surprising given that the majority of immune cells within brain tumors are macrophages either resident or infiltrated from the periphery (27, 29, 30). Recently, astrocytes have been proved to influence a subtype of microglia/ macrophage expressing CD74.

CD74 is among the most upregulated genes in human microglia in the context of brain tumors and other pathologies (31) (Figure 1B). The association of CD74 in microglia/macrophages and brain disorders have been recently extended and validated by scRNAseq approaches comparing healthy and brains affected by autoimmune disorders, neurodegeneration or ischemia. Cd74 upregulation was consistently found in disease-associated macrophages including peripheral macrophages infiltrating the brain, non-parenchymal resident macrophages (meningeal, perivascular, and choroid plexus macrophages) as well as in one subclass of microglia (26, 32) (Figure 1B).

Funtionally, the CD74+ microglia/macrophages were shown to reduce the secretion of IFN-γ in the tumor microenvironment, which would contribute to established an immunosuppressed niche (33). More recently, the increase of CD74+ microglia/macrophages in the context of brain metastases was shown to be dependent on the presence of pSTAT3+ reactive astrocytes, describing a cross-talk between both cell types (19). The ligand of CD74 receptor, MIF, is highly enriched in the secretome of pSTAT3+ reactive astrocytes. CD74+ microglia/macrophages are preferentially located within the metastatic lesion. At this location CD74 could be found translocated in the nucleus where it promotes the expression of NFκβ downstream targets, such as midkine (19), a secreted molecule that accumulates in the extracellular space to promote cell viability (34) (Figure 1C). The importance of MIF binding to CD74+ microglia/macrophages was demonstrated by the reduction of brain metastasis upon treatment with the BBB-permeable MIF inhibitor ibudilast in organotypic cultures (19). Interestingly, ibudilast has been successfully used in patients with multiple sclerosis (35) and in experimental models of glioblastoma (36), which inspired a recently initiated clinical trial (NCT03782415). Although the biology of CD74+ microglia/macrophages remains poorly characterized, its strong association with different brain disorders and its diverse set of functions including the role as a chaperone for the MHCII complex (37), the modulation of migration by interacting with myosin (38) and the activation of NFκβ pathway (34) suggest relevant implications in disease.

T regulatory cells have been described to actively modulate astrocyte behavior in ischemia (42). After stroke there is a massive accumulation of T regulatory cells in the brain that promotes neurological recovery. T regulatory cells are initially attracted to the ischemic brain by CCL1 and CCL20 produced by astrocytes and oligodendrocytes and later expanded by the combined action of IL-2 or IL-33 and T cell receptor recognition. Expanded T regs secrete the EGFR ligand amphiregulin (AREG) that decreases the expression of several astrocyte markers associated with potential negative effects on neuronal viability (42) (Figure 2A). In fact, intraventricular administration of AREG reduced neurological dysfunction associated with Treg-depleted-mice (42). Thus, T regulatory cells contribute to the control of brain damage by modulating astrocyte behavior.

Additional evidences of the crosstalk between T regulatory cells and astrocytes exist in EAE. Administration of an anti-CD3 antibody intranasally in this experimental model was shown to activate IL-10-producing T regulatory cells in the cervical lymph nodes (43). These T regulatory cells moved and became enriched in the CNS where they influence reactive astrocytes expressing the IL-10 receptor subunit alfa (Il10ra). Activation of IL10R1-dependent signaling in astrocytes decreased gene expression patterns typically associated with different aspects of EAE pathophysiology such as BBB degradation (Mmp3, Mmp9), monocyte recruitment (Ccl2), and microglial regulation (Csf2) (Figure 2B). Interestingly, astrocyte-specific downregulation of Il10ra fully impaired the clinical benefit provided by the administration of the anti-CD3 antibody in a pre-clinical model of multiple sclerosis (43).

Primary and secondary brain tumors are infiltrated with CD4+ CD25+ Foxp3+ T regulatory cells (44, 45). In addition, Treg signatures seem to predominate over those related to T-cell function involved in their activation or TCR antigen binding even when measured systemically (46). Consequently, dissecting the biology of the Treg compartment in brain tumors and its cross-talk with other components of the microenvironment including reactive astrocytes might help to develop novel strategies of immunotherapy.

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.