Chromatin network rearrangements in differentiation and cancer reveal connections between evolutionary processes and gene regulation

Abstract

AbstractThe availability of unprecedented datasets detailing the organization of chromatin across species, developmental stages, cell types and diseases invites an integrated analysis of genome organization in three dimensions and its biological significance.In this study we hypothesize that topological properties of the chromatin network can impact plasticity and variability of phenotypes. Our observations lead us to reconsider the theory according to which cancer cells re-acquire pluripotency through reactivation of early evolutionary programs characteristic of species lacking complex body plans or even multicellularity.Recent work has shown that promoter sequence characteristics, including general GC content, TATA-box motifs and number of TF binding motifs, can be predictive of gene expression variability across tissues and individuals in multiple species. Exploiting this dataset and others, we show that genes with specific promoter features and associated similar values of expression variability tend to have specific epigenomic marks and form preferential interactions in 3D, as quantified by chromatin assortativity. We further demonstrate that rearrangements of the 3D topology of the chromatin network in differentiation and oncogenesis can change these associations between gene regulation and evolutionary age. We observed that evolutionary older genes, also broadly corresponding to the least variable genes, are strongly clustered in the chromatin network.We characterize the topology of promoter-centered chromatin networks as well as assortativities of gene ages and expression variability in multiple datasets capturing topological changes in differentiation (embryonic stem cells, B-cells) and find some of these changes to be reversed during oncogenesis.