2.2. In Neuroinflammation
The first detectable interaction between microglia and astrocytes in neuroinflammation concerns the induction of reactivity. About 30 years ago, Matsumoto and colleagues
[43] found that “reactive microgliosis” preceded “reactive astrogliosis” in a mouse model of AD (8–12 weeks old). Based on this temporal sequence, it was assumed that microglia may induce astrocyte reactivity
[1][44]. This relationship was then transposed into the A1/A2 paradigm, suggesting that M1 microglia may trigger the expression of the A1 phenotype of astrocytes
[36]. However, in view of the marked dependence of astrocyte and microglia phenotypic heterogeneity from context conditions, this activation hierarchy seems to contrast the mutual interdependence linking the two cell types in CNS diseases. For example, recent evidence collected in vitro suggests a role of astroglia in the activation of microglial immune response induced by obesity
[32]. To be recalled, the scenario of heterotypic interactions is even more complex and also involves neurons, which may promote microglial reactivity by releasing several signaling molecules such as the chemokines (C-C motif) ligand 2 (CCL2)
[45] and fractalkine
[46][47], and extracellular alarmin high mobility group box protein-1 (HMGB-1)
[48], as well as by increasing extracellular levels of ATP
[49] and glutamate
[50]. Moreover, cytotoxic unfolded peptides on neurons and cell debris may also represent pro-inflammatory factors
[7]. According to their neurotrophic and clearance roles, astrocytes are possibly able to rapidly sense the accumulation of these molecules in the early phases of neurodegenerative diseases, such as AD. Indeed, these cytotoxic molecules have been described to induce fragmentation of astrocyte processes (G), resulting in disruption of astrocyte meshwork in aged rats
[51]. Furthermore, structurally altered proteins activate the NF-κB signaling pathway
[52] and inhibit the astrocyte support to synaptogenesis in prion-infected mice (3–4 weeks old)
[53]. Whether or not these responses may be categorized as signs of astrocyte reactivity, they induce significant changes in the nervous tissue environment. In addition, the disruption of the astrocyte meshwork hampers heterotypic interactions in the aged nervous tissue. In particular, our study in the hippocampus of aged rats (88 weeks old) demonstrated that local interruptions of the astroglial meshwork imply a decrease in their direct interactions with microglia (F) and are responsible for the microglial shift from branched to amoeboid morphology
[33] (C,F), thus providing a rationale of their impaired clearance efficiency. It appears that the impairment of typical astrocyte tasks addressed to nerve cells, such as trophic support and clearance activity may give rise to a context where the onset and progression of neurodegenerative diseases are favored. These data suggest that homeostasis alterations in the CNS may elicit multiple responses from different cell types, thus stressing the idea of neuroinflammation as a choral reaction to pathological stimuli.
It has been, however, assessed that in neuroinflammation, intense molecular crosstalk between glial cells is maintained via a variety of molecules: growth factors, gliotransmitters, cytokines, chemokines, innate-immunity mediators, ATP, mitogenic factors, nitric oxide (NO), ROS and glutamate. Recent evidence collected from a study on 17-week-old mice indicates that microglia promote neuroprotective response from astrocytes by releasing cytokines, such as interleukin-1 beta (IL-1β), Tumor Necrosis Factor alpha (TNF-α) and IL-6
[54]. On the other hand, reactive astrocytes may promote microglial phagocytosis by releasing the complement factor C3
[52] or ATP
[55]. Moreover, astrocytes may either reverberate activation signaling to microglia, thus promoting their migration towards injury sites and phagocytosis, or inhibit their reactivity when physiological homeostasis is restored
[1]. The exchange of extracellular vesicles containing active molecules, such as mRNA fragments is suggested as an additional mechanism of micro-astroglial interaction: in vitro, extracellular vesicles released by microglia, affect astrocyte expression of reactivity markers and the production of components of the extracellular matrix
[56], whereas vesicles produced by astrocytes modulate microglial migration and phagocytosis
[57]. Among these activities, astrocyte production of ECM molecules, such as fibronectin, represents a further strategy of modulating the microglial immune response. It is long known that such molecules may regulate microglial expression of integrin-β1
[58], a mechano-receptor involved in cell–ECM as well as cell–cell mechano-signaling
[59][60]. A direct correlation between integrin-β1 expression and onset of reactivity in microglia was also demonstrated in cultures of murine primary cells
[61] and ex vivo in mice models (8–10 weeks old) of MS
[62]. A study carried out in 3D collagen substrata and organotypic slices assessed the dependence of microglia migration on their expression of integrin-β1
[63]. Our previous studies on direct cell–cell interactions between astrocytes and microglia indicated the recruitment of integrin-β1 at contact sites between their processes, along with focal increases of the concentration of ionized calcium-binding adapter molecule (Iba1)
[33], a microglial marker of Ca
2+ dependent cytoskeletal remodeling, involved in both migration and phagocytosis
[20]. These data suggest that networks of mutual modulations, affecting inflammatory responses of astrocytes and microglia, may be utterly complicated by the integration of molecular and mechanical processes. A paradigmatic example of astrocyte-microglia interactions in neuroinflammation is provided by the occurrence of “triads”, in which astrocytes and microglia cluster with damaged nerve cells and promote clearance of cytotoxic neuronal debris from the nervous tissue
[28][30][64]. In detail, processes originating from an astrocyte form a kind of “mini scar” encircling a terminally injured neuron, split the cell into debris and expose them to microglial phagocytosis.
It is known that long-term activation of the mutual inductions cycle linking microglial and astroglial reactivity may result in a decrease of their activities of maintenance and support, along with direct neurodegenerative effects. As previously stated, the high metabolic activity of reactive astrocytes and microglia may lead to the overproduction of ROS, which in turn, induces oxidative stress and neurodegeneration in the nervous tissue
[12][65]. On the other hand, neurodegenerative effects of oxidative stress may be counterbalanced by the neuroprotective activity of NO
[66]: in the hippocampus of adult rats, it has been demonstrated that astroglial production of NO induced the expression of heme-oxygenase-1, an enzyme involved in the synthesis of the known antioxidant bilirubin
[67]. Moreover, the production of NO by microglia and astrocytes may trigger clearance of damaged cell debris and defense against invading bacteria
[68]. Nonetheless, NO may inhibit axonal conduction and induce neurological disorders in adult rats
[69] and enhance cytotoxicity mediated by NMDAR
[12].
A high concentration of cytokines involved in astrocyte-microglia crosstalk during inflammation may also induce neurodegeneration. Evidence has been provided that injection of IL-1 in ischemic rats resulted in neurodegenerative effects
[70]. Conversely, a study in mouse models of cerebral ischemia demonstrated that overexpression of IL-1 through viral transfection exerted beneficial effects
[71]. Neurodegenerative induction by cytokines was also observed in transgenic mouse models overexpressing IL-6 (age: 12, 24 and 48 weeks)
[72] and TNF-α (age: 11 weeks)
[73]. Since these molecules are proinflammatory, it is not surprising that their neurodegenerative effects were found to involve other modulators of inflammation. This is the case of IL-1
[74] and TNF-α
[75], which were demonstrated to activate NO production in co-cultured human astrocytes and neurons, whereas synergy between IL-6 and transforming growth factor β (TGFβ) generated pathogenic T
H17 lymphocytes in in vitro and ex vivo experiments on mice
[76]. Changes in the ECM may also be involved in the activation and, eventually, intensity regulation of the immune response. Evidence showed that expression of fibronectin by astrocytes may be increased following seizures induced by kainic acid (in 12-week-old rats)
[77]. On the other hand, a study performed on animal models of spinal cord injury (12-week-old mice) indicated that increased expression of astroglial fibronectin may induce the exacerbation of the immune response
[78]. These data seem to suggest that astrocyte and microglia during inflammation activate a cycle of self-triggering, mutual inductions, which coordinate but also steadily amplify their responses until either resolution of the disease or onset of new neurodegenerative processes. However, growing evidence indicates that non-inflammatory processes modulating synaptic connectivity are finely tuned by cytokines during development and in health. A representative example is provided by IL-33, a member of the IL-1 family that showed either neuroprotective or neurodegenerative effects in different experimental conditions. Intraperitoneal injections (50 mg/kg, about 1.15 mg per animal, 10 days) followed by intrahippocampal injection (400 ng by side) of IL-33 in 8-week-old mice, evoked neuroinflammation and cognitive impairment
[79], whereas intraperitoneal injections (200 ng, 7 days) of IL-33 in transgenic mouse models of AD (mice, 48 weeks old) mobilized microglia to prevent and clear Aβ-deposits, thus ameliorating cognitive impairment
[80]. Of note, IL-33 released by astrocytes has been suggested to modulate typical microglia activities such as the pruning of synapses in mouse embryos during maturation
[81] and modulation of synaptic plasticity in adult mice (16 weeks old)
[82]. In cell cultures, TNF-α has been demonstrated to promote astrogenesis from human neural progenitors
[83] and the maturation of human neuroblasts
[84]. These data indicate that cytokines may be involved, either as effectors or mediators, in dynamic lifespan changes of the CNS. In neuroinflammation, a massive increase of cytokine concentration causes a sudden, choral activation of neuroinflammatory processes and, at the same time, triggers inherent, mutually reinforcing counter effects. Indeed, a metabolic acceleration involves both astrocytes and microglia, promoting oxidative stress and, therefore, cell damage. The increasing cytotoxicity induced by these noxious counter effects is constantly weighted against neuroprotective effects of inflammation and, in the course of time, may result in a neurodegenerative outcome. Finally, in the aged CNS, a condition of chronic inflammation involving dysregulation of astrocyte-microglia interactions favors the onset of age-related neurodegenerative diseases.