Genomic analysis of highly purified astrocytes reveals in vivo astrocyte gene expression: A new resource for understanding astrocyte development and function

Although astrocytes constitute nearly half of human brain cells, their functions are one of the great unsolved mysteries of neurobiology. Here I describe the first method for the isolation and purification of developing and mature astrocytes from mouse forebrain. This method takes advantage of the expression of S100beta by astrocytes. I used fluorescent activated cell sorting (FACS) to isolate enhanced green fluorescent protein (EGFP) positive cells from transgenic mice that express EGFP under the control of an S100beta promoter. Depletion of astrocytes and oligodendrocytes enabled me to obtain purified populations of neurons, while panning with oligodendrocyte-specific antibodies enabled me to obtain purified populations of oligodendrocytes. Using Affymetrix GeneChip Arrays I then created a transcriptome database of the expression levels of over 20,000 genes by gene profiling the mRNA isolated from these three main CNS cell types. This database provides the first global characterization of the genes expressed by mammalian astrocytes in vivo and is the first look at the astrocyte transcriptome. This data reveals many novel aspects of astrocyte phenotype, including a large degree of metabolic specialization and evidence for a novel role of astrocytes as professional phagocytes. By comparing different stages of astrocyte and oligodendrocyte development, I have identified new candidate genes involved in glial development, including the first astrocyte-specific transcription factors. In addition, I demonstrate that Aldh1L1, a highly expressed astrocyte-specific gene, is a highly specific antigenic marker for astrocytes with a substantially broader, and therefore potentially more useful, pattern of astrocyte expression than the traditional astrocyte marker GFAP. My findings also call into question the concept of a "filial" cell class, as the gene profiles of astrocytes and oligodendrocytes are as dissimilar to each other as each of they are to neurons. These astrocyte, neuron, and oligodendrocyte transcriptome databases will serve as a resource to the entire neuroscience community by providing improved cell type specific markers and for better understanding of neural development, function, identifying new drug targets, and predicting mechanisms of pathology in human disease.