Introduction

Phagocytosis is an important process in the immune system. Exogenous viruses and pathogens as well as endogenous cell debris and abnormally aggregated proteins are removed by phagocytes, "professional" phagocytic cells that have higher phagocytic capability than other cells. In the brain, microglia, sometimes referred to as tissue-resident macrophages, are thought to be the major immune and phagocytic cells; thus, their phagocytic activity has been well studied. However, recent studies have shown that astrocytes are also capable of phagocytosis.¹ Astrocytes are glial cells with a wide variety of functions, including taking up neurotransmitters, supplying nutritional factors to neurons, and forming the blood-brain barrier. Astrocytic phagocytosis, which occurs in parallel with these diverse functions, differs in many respects from microglial phagocytosis. Here, we will discuss phagocytosis in the brain by comparing the differences between microglia and astrocytes.

Phagocytosis of apoptotic cells

Damisah et al. compared phagocytosis of a single apoptotic neuron by microglia and astrocytes.² The authors induced apoptosis at the single-cell level using laser irradiation and observed the response of surrounding glial cells by in vivo two-photon imaging. They found that microglia engulfed the cell body and proximal dendrites, while astrocytes phagocytosed more distal dendritic processes. The authors propose that the "eat me" signal presented by apoptotic cells may be different between the cell body and the distal processes. They further reported that in the absence of microglia, astrocytes ingested and removed the cell bodies instead, although it took approximately 50 hours for them to remove apoptotic bodies, which was twice as long as the time required for removal by microglia.

Lööv et al. studied lysosomal pH during phagocytosis using primary cultured cells.^3,4 To acquire temporal information on degradation and lysosomal pH, dead cells labeled with pH-sensitive dyes (which fluoresce under acidic conditions) were added to spleen macrophages or primary cultured astrocytes. Time-lapse imaging revealed that in spleen macrophages, fluorescence indicating uptake of dead cells into the acidic lysosome was observed after 3 hours, and degradation was confirmed by the disappearance of the aggregated nuclei of dead cells after 3 days. In contrast, the nuclei of dead cells in astrocytes disappeared after 12 days, and pH-sensitive fluorescence was hardly observed before this time point. These observations suggest that astrocytes take longer to degrade dead cells than spleen macrophages and that the lysosomal pH of astrocytes is higher than that of spleen macrophages. The authors hypothesized that a higher pH in astrocytic lysosomes causes slower degradation. To test this hypothesis, astrocytes were treated with microparticles that cause lysosomal acidification. The results show that degradation was accelerated after 5 days. Although the reason that the pH of lysosomes is higher in astrocytes than in spleen macrophages is unclear, this difference in pH may cause a temporal difference in the degradation of dead cells.

Microglia and astrocytes not only share a role in phagocytosis. An intriguing recent report by Konishi and colleagues showed that when microglia were genetically depleted, astrocytes phagocytosed and removed microglial debris within approximately 4 days.⁵ In addition, when the phagocytic function of microglia was impaired, astrocytes instead phagocytosed dead cells, which is consistent with reports by Damisah et al. These results suggest that astrocytes may compensate for the failure of microglial phagocytosis.

Phagocytosis under pathological conditions

Microglial and astrocytic phagocytosis under pathological conditions were directly compared by Morizawa et al.⁶ In their report, the researchers induced cerebral ischemia in the mouse brain by middle cerebral artery occlusion (MCAO) and revealed spatiotemporal differences in phagocytosis by microglia and astrocytes using immunohistochemical analysis. Microglia were found in both the ischemic core and penumbra region (with reduced blood flow but no cell death), phagocytosing cell debris of various sizes. In contrast, astrocytes were found only in the penumbra region and phagocytosed relatively small (less than 10 μm²) cell debris. Furthermore, the temporal change in the expression of phagocytic markers was also different between microglia and astrocytes. In microglia, the expression of Galectin-3 (a lectin involved in phagocytic function) and CD68 (a microglial lysosomal marker) peaked at 3 days after MCAO and returned to control levels after 14 days. In astrocytes, the expression of Galectin-3 and LAMP1 (astrocytic lysosomal marker) peaked at 7 days and was still elevated at 14 days after MCAO. This astrocytic phagocytosis is required for ATP-binding cassette transporter A1 (ABCA1), the expression of which is upregulated in an astrocyte-specific manner during ischemia. The absence of upregulated ABCA1 expression in microglia suggests that phagocytosis by astrocytes and that by microglia may involve different molecular mechanisms.

Phagocytosis of synapses

In the dorsal lateral geniculate nucleus (dLGN), which receives projections from the optic nerve, excess synapses undergo synaptic pruning during development. Two groups have reported that both microglia and astrocytes are involved in synaptic pruning via different receptors.^7,8 Astrocytes phagocytose synapses via Mer tyrosine kinase (MERTK) and multiple EGF-like-domains 10 (MEGF10), while microglia phagocytose synapses via C3-CR3 signaling, a complement pathway. However, phagocytosis of synapses by both microglia and astrocytes is regulated in a neural activity-dependent manner. In addition, as for microglia, several other signals involving phagocytosis of synapses have been reported: CD47-signal regulatory protein α (SIRPα) signals acts as a "do not eat me" signals and sushi repeat protein X-linked 2 (SRPX2) specifically suppresses phagocytosis of VGLUT2⁺ synapses.^9,10 Collectively, these findings suggest that microglial synaptic phagocytosis is mediated by a variety of signaling pathways, which may be responsible for the precise differentiation of synapses to be eliminated. However, few reports have investigated astrocytic synapse elimination, and more detailed verification is required.

Discussion of role sharing in phagocytosis between microglia and astrocytes

Under healthy and pathological conditions, astrocytes tend to phagocytose relatively smaller cell debris than microglia.^2,6 In addition, astrocytes take longer than microglia to engulf and degrade debris, which may be caused by a high lysosomal pH.^3,4 The differences in phagocytic properties between microglia and astrocytes may partly result from the differences in the origin of each cell type. Astrocytes, like other cells in the brain, such as neurons, are derived from the ectoderm, whereas microglia are thought to be derived from the mesoderm, and the basic characteristics of cells differ according to their origin.

Astrocytes are supportive cells that are important for the maintenance of the physical structure of the brain and occupy basically nonoverlapping territories.¹¹ Moreover, given their various roles in assisting neuronal function that are supported by their relatively stable microstructures, such as tripartite synapses, it is reasonable for astrocytes to phagocytose cells to the extent that they do not cause significant structural changes to themselves. Microglia, on the other hand, are highly dynamic cells that continuously move their branched processes. In addition, microglia express a large number of receptors so that they can rapidly sense abnormalities in the extracellular milieu, such as cell death or viral infections, and can drastically migrate toward damaged brain areas. Furthermore, during phagocytosis, microglia are able to retract their processes to adopt an ameboid shape, making them suitable for phagocytosis of large structures. Thus, the role sharing between microglia and astrocytes in phagocytosis described above may be a result of the innate characteristics of each cell type. On the other hand, in the absence of microglia themselves or microglial phagocytic function, astrocytes phagocytose more and larger cell debris than they normally phagocytose. Although the mechanism underlying the increased phagocytic capacity of astrocytes is not clear, it has been suggested that microglia may inhibit astrocyte phagocytosis to some extent under normal conditions.^2,5 In any case, it would be beneficial for brain homeostasis for astrocytes to compensate for the phagocytic function of microglia.

With respect to differences in time scales, it has been hypothesized that the phagocytic function of astrocytes increases compensatory function only when microglial phagocytosis reaches its limit.¹² Damisah et al. showed that even in the absence of microglia, astrocytes are capable of phagocytosing larger debris than usual, although the time required for phagocytosis by astrocytes is longer than that required for phagocytosis by microglia. These findings suggest that the slowness of phagocytosis by astrocytes may not be due to a microglia-dependent mechanism but to intrinsic differences in the molecular mechanisms that regulate phagocytosis. Lööv et al. suggested that the pH of lysosomes in astrocytes is maintained at a high level to inhibit antigen degradation, which may be beneficial for the presentation of antigens by astrocytes.³ However, whether astrocytes have antigen-presenting functions remains unclear. To elucidate the role sharing of microglia and astrocytes in phagocytosis in detail, it is necessary to identify the molecular mechanisms of each process of phagocytosis, such as receptor activation, uptake, lysosome formation and degradation of phagocytosed materials in both cell types.

Whether a synapse is phagocytosed by microglia or astrocytes might be determined by the nature of each synapse. As mentioned above, it has recently been shown that synapse elimination by microglia is tightly regulated by multiple factors, which might be major players in synapse elimination by microglia. In addition to microglia, the cellular and molecular mechanisms by which astrocytes regulate synapse elimination need to be elucidated. For this purpose, it is necessary to establish an experimental system for simultaneous imaging of synaptic phagocytosis by astrocytes and microglia.

Summary

In summary, few reports have compared the phagocytic capacities of microglia and astrocytes simultaneously under the same conditions. Both synapses and apoptotic cells as well as amyloid-β and myelin are phagocytosed by both astrocytes and microglia.^13-16 The mechanism by which other targets are phagocytosed by these cells should also be investigated. Finally, cell type-specific genetic manipulation will enable us to simultaneously and directly compare phagocytosis by microglia and astrocytes, and it would shed light on role sharing between them in phagocytosis.

References

1. Cahoy, J. D. et al. : J. Neurosci., 28 (1), 264 (2008).
2. Damisah, E. C. et al . : Sci. Adv., 6 (26), eaba3239 (2020).
3. Lööv, C. et al. : Glia, 63 (11), 1997 (2015).
4. Lööv, C. et al. : PLoS One., 7 (3), e33090 (2012).
5. Konishi, H. et al. : EMBO J., e104464 (2020).
6. Morizawa, Y. M. et al. : Nat. Commun., 8 (1), 28 (2017).
7. Schafer, D. P. et al. : Neuron, 74 (4), 691 (2012).
8. Chung, W. S. et al. : Nature, 504 (7480), 394 (2013).
9. Lehrman, E. K. et al. : Neuron, 100 (1), 120 (2018).
10. Cong, Q. et al. : Nat. Neurosci., 23 (9), 1067 (2020).
11. Halassa, M. M. et al. : J. Neurosci., 27 (24), 6473 (2007).
12. Magnus, T. et al . : J. Neuropathol. Exp. Neurol., 61 (9), 760 (2002).
13. Wyss-Coray, T. et al. : Nat. Med., 9 (4), 453 (2003).
14. Mills, E. A. et al . : Proc. Natl. Acad. Sci. USA., 112 (33), 10509 (2015).
15. Heckmann, B. L. et al. : Cell, 178 (3), 536 (2019).
16. Hughes, A. N. et al. : Nat. Neurosci., 23 (9), 1055 (2020).