Mechanisms of naturally-occurring astrocyte death during retinal development

Duke University, Durham NC

Investigators

Linked publications & trials

Abstract

Naturally-occurring cell death is essential to pattern formation in the developing nervous system. Retinal astro- cytes undergo an unusual non-apoptotic form of death mediated by microglia; this mechanism is crucial for pat- terning astrocytes into a template that controls angiogenesis. Despite this importance, the mechanism by which microglia kill astrocytes is unknown. To unravel these mechanisms one must know the cellular and mo- lecular characteristics of the microglia that kill astrocytes, but this too remains unknown. The objective here is to identify microglia that execute developmental astrocyte death and the mechanisms that enable this function. The central hypothesis is that a developmentally transient microglial population, termed nerve fiber layer- associated microglia (NFL-M), responds to astrocyte-derived “eat-me” signals and mediates astrocyte elimina- tion. The rationale for this work is that knowledge of astrocyte death mechanisms will make it possible to learn how the astrocyte and vascular networks arrive at their mature pattern – in both normal and pathological de- velopment. Further, as non-apoptotic “death by phagocyte” can occur also in neurodegenerative disease, in- sight into death mechanisms may have broad relevance for disease pathobiology. The Specific Aims are: 1) Identify retinal microglial subsets that are recruited by astrocyte “eat-me” signals. Preliminary studies show that phosphatidylserine (PtdSer) is an astrocyte cell-surface “eat-me” cue that recruits phagocytic micro- glia. Single-cell RNA-sequencing identified a microglial population, the NFL-M, that is likely to be the astrocyte- responsive subset. This working hypothesis will be tested by manipulating astrocyte density and PtdSer expo- sure, which should alter the number and/or retinal distribution of NFL-M if the hypothesis is correct. 2) Identify microglial populations that execute astrocyte engulfment. Prolonged bouts of phagocytosis require high mitochondrial output. NFL-M show transcriptional signatures of intense mitochondrial metabolism including ac- tivation of Nrf2 signaling, a key response to metabolic stress. These data suggest NFL-M are highly phagocytic towards astrocytes; this hypothesis will be tested in vivo using genetic manipulations that prevent microglia from assuming the NFL-M state or activating Nrf2. 3) Identify genes expressed by astrocyte-associated microglia that mediate their function. If astrocytes are killed by a transcriptionally unique microglial popula- tion, the genes that make them unique should support this function. Lgals3, a gene with known roles in phago- cytosis, is selectively expressed by NFL-M. Preliminary data show that it is crucial for microglial phagocytosis in retinal degeneration, supporting the working hypothesis that it will also be required for astrocyte removal. This will be tested using Lgals3 conditional mutant mice. Completion of this work is expected to define specific microglial subsets that kill developing astrocytes and their contribution to retinal astrocyte and vascular pattern- ing. This will be significant because it will open the way to defining mechanisms underlying death by phagocyte and the contributions of this novel death mechanism to nervous system development and disease.