Mapping tissue development and heterogeneity
in health and disease with signalling entropy

Christopher Banerji

Time: 12:30 - 1:10 PM

Type: Invited presentation

Affiliation: Centre for Mathematics and Physics in the Life Sciences and Experimental Biology (CoMPLEX), University College London, UK

Abstract

Cellular differentiation is a fundamental biological process, essential for the development and maintenance of multicellular life; conversely, the subversion of this process is the foundation for some of the most devastating human pathologies, notably cancer. Recently, certain global principles of what characterises pluripotent stem cell populations have emerged, positing a strong role for gene expression heterogeneity. We postulated that signalling entropy, a single sample, network theoretic measure of interactome signalling promiscuity and intra-sample heterogeneity, derived from genome wide gene expression data, may prove a unifying quantification of a cells position in the global differentiation hierarchy. By analysing over 1,000 healthy tissue samples we demonstrated that signalling entropy correlates with cell potency across multiple lineages, being highest in embryonic stem cells and decreasing systematically over differentiation time courses. We also revealed that our measure is elevated in cancerous as opposed to healthy tissue, and in cancer stem cells as opposed to the tumour bulk. We next considered our measure as a prognostic indicator in epithelial cancer, analysing over 5,000 primary tumour samples. We demonstrated the signalling entropy is strongly prognostic in both breast and lung adenocarcinoma, out-performing current prognostic indicators. Our measure is found to be a robustly prognostic, valid regardless of oestrogen-receptor status in breast cancer and within the stage I stratum in lung adenocarcinoma. We thus present signalling entropy as a means to chart the differentiation landscape, providing insight into regenerative medicine and disorders of development.