Deep proteomic analysis of human microglia and model systems reveal fundamental biological differences of in vitro and ex vivo cells

Abstract

Using high resolution quantitative mass spectrometry, we have generated the most comprehensive human and mouse microglia proteomic datasets to date, consisting of over 11,000 proteins across all six microglia groups. Microglia from different sources share a core protein signature of over 5600 proteins, yet fundamental differences are observed between species and culture conditions, indicating limitations for human disease modelling in mouse or in in vitro cultures of microglia. Mouse ex vivo microglia show important differences at the proteome level such as differential expression of inflammation and Alzheimers Disease associated proteins. We identify a tenfold difference in the protein content of ex vivo and in vitro cells and significant proteome differences associated with protein synthesis, metabolism, microglia marker expression and environmental sensors. Culturing microglia induces rapidly increased growth, protein content and inflammatory protein expression. These changes can be restored by engrafting in vitro cells into the brain, with xenografted hESC-derived microglia closely resembling microglia from human brain. This data provides an important resource for the field and highlights important considerations needed when using model systems to study human physiology and pathology of microglia.