Disentangling Sources of Phenotypic Variability in Oncogene-addicted Cancers

Abstract

Tumor heterogeneity and plasticity are key contributors to treatment insensitivity and eventual recurrence in oncogene-addicted cancers. Sources of this variability have been ascribed to genetic, epigenetic, stochastic, and multiple other sources. Multi-omics datasets, single-cell lineage tracing, and quantitative drug response experiments were performed on multiple in vitro models prior to and during the course of targeted therapy treatment to assign sources of tumor variability to differential treatment outcomes. In a study dedicated to tumor heterogeneity, we identified that EGFR-mutant lung cancer cell line ‘versions’, evolved independently in culture at different institutions, exhibited genetic differences, while single-cell derived sublines displayed predominately epigenetic differences, leading to variable drug response profiles. A single subline showed clear functional genetic differences, which led to resistance. A study of tumor plasticity identified a novel drug-tolerant ‘idling’ population state in the response of BRAF-mutant melanoma cell lines to treatment. Lineage tracing coupled to single-cell gene expression provided evidence for a mathematical model-predicted mechanism of epigenetic state transitions, by which drug modifies the phenotypic landscape, resulting in cell population re-equilibration to the modified landscape. Transcriptomic and epigenomic profiling pointed towards ion channel activity as being characteristic of the drug-modified ‘idling’ landscape, which was subsequently targeted by a form of targeted cell death. Together, these studies support for a multi-tier view of tumor variability, where a complex network of sources contribute to variable treatment outcomes and evasion. Furthermore, these results advocate for novel therapeutic strategies that target network control points in conjunction with traditional targeted therapies.