A dynamical systems approach to understand plasticity and non-genetic heterogeneity in cancer

This proposal addresses non-genetic heterogeneity in ER+ breast cancer (BC), a cancer with high rate of relapse.

The overarching hypothesis is that cell plasticity underlies non-genetic heterogeneity and allows quiescence or treatment resistance, to emerge.

We hypothesise that plasticity, defined as the ability to reversibly switch cellular states, is due to gene expression oscillations from combined transcriptional and post-transcriptional mechanisms.

This hypothesis builds on emerging evidence that oscillations represent a prevalent, yet under-appreciated and under-studied mode of regulation, that is inherently plastic and specifically enables cell-state transitions.

Oscillations are a hallmark of a dynamical system that is not at equilibrium; such systems cannot be understood by biological experimentation and intuition alone.

Therefore, we will combine experimentation and mathematical theory in an iterative, integrated, and bidirectional way to address the 3 aims of this proposal. Aim1. Quantitative characterisation of Hes1 dynamical gene expression during cancer cell state transitions. Aim2.

Developing computational modelling and Bayesian inference to understand how changes in oscillatory gene expression occur. Aim3.

Experimentally manipulating dynamics to test their functional significance for cell state transitions In Aim 1, using ER+ BC cells and CRISPR/Cas9 knocked-in fluorescent reporter, fused in frame with the endogenous protein and new live reporter lines for BC stem cells, we will analyse the oscillatory expression of Hes1, a key transcriptional regulator in cell state transitions, by live-imaging and at the single-cell level.

To understand how single cell oscillatory dynamics are integrated and influenced by the environment, we will test the effect of 3D organisation and added factors. We will validate our findings in several cancer lines and human primary tumour samples.

In Aim 2, we will develop stochastic mathematical models with delay, couple with quantitative experimental data, to understand the influence and likely origins of gene/protein expression noise in transcription/translation networks and Bayesian inference methods to predict change sin parameters that drive changes in dynamics during cell state transitions.

In Aim 3, we will test emergent hypotheses for the mechanism of cell state transitions, by manipulating protein expression dynamics by optogenetics, controlled degradation and other chemical methods.

We envision using this knowledge in the future for designing treatments that would “tune” the expression of dynamically expressed proteins rather than repressing or overexpressing them.

For example, one may screen for modifiers of Hes1 dynamics as a versatile tool to suppress plasticity during treatment or to lock cancer cells into quiescence later, reducing the risk of relapse.