Quantitative measurement of phenotype dynamics during cancer drug resistance evolution using genetic barcoding

Introduction

The evolution of resistance remains the primary obstacle to the successful treatment of cancers. Initial studies focussed on resistance as a consequence of genetic alterations that produce resistant phenotypes that drive clonal evolution^{1, 2–3}. Recent focus has shifted to nongenetic mechanisms of resistance (often termed ‘plasticity’) that enable cells to rapidly change phenotype^{4, 5–6} promoting adaptive evolution following the change in selective pressure imposed by treatment. Efforts to tackle treatment resistance require knowledge of the molecular mechanism responsible so that it can be targeted. For example, new generations of targeted drugs use the resistance mechanisms of previous drugs as novel targets⁷. However, evolutionary-informed treatment strategies are also reliant on an understanding of the behaviour of resistant phenotypes: approaches to “steer” tumour evolution to forestall or prevent resistance emergence require an understanding of the heritability of resistance mechanism(s) through cell divisions⁸, and the relative fitness of resistant cells compared to drug sensitive counterparts⁹.

Resistance evolution offers a unique opportunity to study phenotypic evolution in cancer: treatment is one of the few environmental changes a tumour experiences with known timing that is under clinical control, and resistance is clearly defined as a phenotype that survives treatment. These features mean studies into resistance are well placed to distinguish evolutionary behaviours such as the stability of a resistance phenotype through cell divisions and the environmental dependence of phenotypic change^{10, 11–12}. These behaviours dictate how resistance is distributed amongst cell lineages over time, which can in turn determine the relatedness of surviving cells.

In a patient tumour, naturally occurring somatic mutations can be interpreted as genetic barcodes¹³. In an experimental setting, lineage tracing technologies enable the tracking of cell relatedness¹⁴: unique genetic sequences are incorporated into cells’ genomes via lentivirus infection, meaning all subsequent ancestors of the parental, barcoded population’s cells inherit this experimentally measurable tag. Quantitative assessment of barcode data enables the measurement of the clonal dynamics¹⁵. When populations of cells are exposed to a change in extrinsic selection pressures, these genetic lineage tracing data provide a means to understand phenotype dynamics. For example, whether dominant lineages are shared between related experimental populations following exposure to drug treatment has been used...