System-level regulation of hierarchical transitions in a tumour lineage

The fundamental principles defining how cell state transitions are regulated along a tumour lineage to determine its cellular composition remain unclear. Here, we investigate how such transitions are controlled in a reductionist hierarchical brain tumour model. Quantitative analysis of 3D transition maps revealed that the differentiation of cancer stem cells (CSCs) into transit amplifying progenitors (TAPs) depends on the identity of their immediate neighbours. This is embodied in a transition rule that quantitatively predicts the probability to differentiate as a function of the proportion of TAP neighbours. Integrated into 3D simulations of tumour growth, this rule spontaneously recapitulates spatial segregation of CSCs in clusters and their stable proportion. We further show in vivo that CSC clustering protects the CSC pool from depletion driven by TAPs. We identify an EGFR-mediated relay mechanism that propagates the CSC-to-TAP transition across CSC clusters. In CSCs, the LRIG1-like EGFR inhibitor Kekkon1 dampens this propagation, ensuring continuous replenishment of the CSC pool. Collectively, these findings show how local fate regulation drives emergent segregation which in turn constrains CSC differentiation dynamics. This provides a conceptual framework to understand and exploit the tumours intrinsic differentiation potential by manipulating the determinants of its system-level features.