Molecular and functional properties of PFC astrocytes during neuroinflammation-induced anhedonia

Abstract

Astrocytes are widely implicated in CNS diseases, but their contributions to disease related phenotypes remain incompletely explored. Anhedonia accompanies several neurological and psychiatric diseases, including major depressive disorder (MDD) and Alzheimer’s disease (AD), both of which are associated with neuroinflammation. In order to explore how neuroinflammation affects astrocytes, we assessed medial prefrontal cortex (PFC) and visual cortex (VCX) astrocytic gene expression using a neuroinflammation mouse model that displayed anhedonia as a phenotype. In this model, anhedonia was reversed by the fast acting antidepressant ketamine. Astrocyte specific gene expression alterations included those related to immune cell signaling, intracellular Ca^2+^ signaling, cholesterol biosynthesis, and metabolic pathways. Such changes peaked when anhedonia was greatest, and reversed to normal when anhedonia subsided. However, region-specific molecular identities between PFC and VCX astrocytes were maintained throughout, implying that astrocyte identities do not converge during neuroinflammation. We also mapped anhedonia-related astrocyte and bulk tissue gene expression changes onto published PFC single cell RNA sequencing data, and compared them to MDD and AD post-mortem human tissue samples to identify shared mechanisms. Finally, we assessed how neuroinflammation affected mPFC neuronal properties and detected no alterations at a time point when there was strong astrocyte reactivity. Our data show that neuroinflammation can cause significant and reversible changes in astrocyte gene expression and mouse behaviour without obvious neurotoxicity or loss of essential homeostatic functions. Furthermore, gene expression signatures accompanying neuroinflammation reveal pathways shared with MDD and AD, which display neuroinflammation as a comorbidity in humans. Astrocytes are widely implicated in brain diseases, but their contributions to disease-related phenotypes remain incompletely explored.