Introduction
Microglia support brain health and resilience during various diseases. Compelling evidence supports a role for a reactive cell state often referred to as disease-associated microglia (DAM) in disease resilience, notably in Alzheimer’s disease and in brain repair after injury. DAMs downregulate homeostatic features and upregulate markers of lysosomal, lipid metabolic and phagocytic pathways. These cells are observed alongside proliferating microglia in many brain pathologies, including acute injuries (e.g., stroke) where synchronised microglial activation facilitates better temporal study of cell state emergence and transitions. Although specific reactive microglial states are clinical targets, it is unclear how they emerge and evolve during disease. Addressing this knowledge gap by temporally tracking transcriptional/functional phenotypes of microglia is crucial for identifying and timing optimal therapeutic approaches.
Aims
To determine if cell-cycle influences emergence of reactive microglial states and their temporal evolution.
Methods
To determine the spatiotemporal distribution of microglial proliferation, photothrombotic stroke was induced on Mki67RFP mice. Tissue was processed for flow cytometry (1-7d post-stroke, n=4) or histology (2-3d post-stroke, n=4). To fate-map proliferating cells, Mki67IRES-CreERT2;Rosa26tdTomato mice underwent sham or stroke surgeries and tamoxifen administered 1-3d post-surgery. TdTomato positive and negative microglial populations were separately sorted at 5d or 28d post-surgery for scRNAseq using 10x Genomics with feature barcoding technology (n=4/group). Brain tissue was taken from a separate cohort to determine the spatial distribution of tdTomato+ microglia at 5d, 14d and 28d post-sham/stroke (n=4). Sex-balanced cohorts were used throughout. All procedures complied with institutional and UK Home Office ethical regulations.
Results
We demonstrate microglial proliferation peaks between 2-3d in a mouse model of stroke. Using our fate-mapping model to time-stamp emerging proliferating microglia, we show these cells exit cell-cycle and daughter cells remain detectable up to 28d post-injury, revealing their chronic persistence. By combining our fate-mapping model with scRNAseq at early (5d) and late (28d) timepoints, we show cell-cycle favours emergence of metabolically active DAM-like cells among other states in the reactive microglial landscape at 5d. By 28d, many of these previously reactive cells return to a homeostatic state, whereas others transition to a chronic DAM-like phenotype marked by expression of antigen presentation pathways. The latter supports that microglial states have plasticity and may functionally (mal)adapt over the course of disease. We are currently applying ex-vivo functional assays to compare the temporal association of transcriptional and functional phenotypes and understand if states are governed by environmental or cell-autonomous signals.
Conclusion
Our work supports that cell-cycle influences reactive microglial diversity by promoting adoption of restricted cell states. Assessing whether cell-cycle is necessary and/or sufficient to become DAM-like is now important for developing new therapies modulating DAMs in human brain disease.