MPTH-29CONNECTING MUTANT GENOTYPES TO ABERRANT TRANSCRIPTIONAL SIGNATURES ACROSS SERIAL SECTIONS OF A HUMAN TUMOR

Identifying the transcriptional consequences of mutations that cause tumors may reveal new targets for the development of cancer therapies. However, this task is complicated by the variability that characterizes tumors from different individuals, which arise from unique genetic backgrounds, consist of various cell types in different proportions, and often exhibit clonal heterogeneity. Here we present a novel strategy to connect mutant genotypes to aberrant transcriptional signatures within individual human tumors. This strategy deconstructs a tumor by extracting RNA from serial sections and comparing genome-wide gene coexpression relationships to control samples analyzed in a similar fashion. In parallel, quantitative genomic analysis of the same serial sections enables identification and mapping of clonal populations within the tumor. We apply this strategy to a single grade II human astrocytoma and identify tumor-specific transcriptional signatures. Genomic analysis reveals clonal heterogeneity, including a dominant clone marked by mutations in IDH1 and TP53. Expression patterns of the two largest tumor-specific transcriptional signatures are almost perfectly associated with the spatial distribution of the dominant clone and consist of genes that are activated or repressed within malignant cells and their microenvironment. Our experimental strategy offers an unbiased and personalized method for identifying aberrant transcriptional signatures that are associated with specific mutant genotypes in any solid tumor while preserving the complex cellular milieu of the tumor microenvironment. More generally, this strategy can reveal shared and distinct patterns of transcriptional organization between any pair of tissue samples, thereby facilitating novel comparisons of diverse biological systems in health and disease.