Abstract
The brain is a major site of relapse for several cancers, yet deciphering the mechanisms of brain metastasis remains a challenge because of the complexity of the brain tumor microenvironment (TME). To define the molecular landscape of brain metastasis from intact tissue in vivo, we employ an RNA-sequencing-based approach, which leverages the transcriptome of xenografts and distinguishes tumor cell and stromal gene expression with improved sensitivity and accuracy. Our data reveal shifts in epithelial and neuronal-like lineage programs in malignant cells as they adapt to the brain TME and the reciprocal neuroinflammatory response of the stroma. We identify several transcriptional hallmarks of metastasis that are specific to particular regions of the brain, induced across multiple tumor types, and confirmed in syngeneic models and patient biopsies. These data may serve as a resource for exploring mechanisms of TME co-adaptation within, as well as across, different subtypes of brain metastasis.