A transcriptomic microglia taxonomy across mouse and human pathologies

Microglia are the key immune cells involved in virtually all diseases of the central nervous system (CNS). Numerous previous single-cell sequencing approaches to the study of microglia have revealed extensive diversity in this cell type in the human and mouse brain in different conditions. Importantly, a unifying, comprehensive, disease- and species spanning transcriptomic immune cell-focused atlas of the perturbed CNS, including true single-cell spatial transcriptomic information on neighborhood relationships, is lacking. Here, we generated a novel large-scale single-cell RNA-sequencing-derived taxonomy through the analysis of > 1 million CNS cells enriched for myeloid cells across > 30 different pathologies and conditions. Disease-associated human microglia, together with CNS-associated macrophages and monocytes, were differentiated iteratively into 27 superclusters and 192 clusters, showing substantial overlap with their counterparts in the corresponding murine pathologies. This top-down approach, in combination with targeted and genome-wide subcellular spatial transcriptomics, enabled us to examine and compare the spatial interactome of the identified superclusters and clusters within and across pathologies and species. Using cell-specific mutants, we further characterized the underlying transcriptional programs of activation-associated microglial clusters, identifying distinct dependencies on type I and II interferons and colony stimulating factor 1 receptor pathways. Our data provide new insights into the spatial dynamics of the endogenous CNS immune system during development, health and disease in human and mouse.