Cancer's ability to evolve in response to the pressures exerted by therapy is the single most common cause of therapeutic failure and poor clinical outcomes, across all cancer types. Despite the definitive link of cancer evolution to poor patient outcomes, we still lack a systemic framework to rationally design therapeutic strategies to anticipate and counter this evolution. We studied the evolution of chronic lymphocytic leukemia (CLL), which epitomizes the challenge posed to modern oncology by cancer evolution: despite effective therapies, the disease invariably recurs. We have developed novel methods to measure the critical dimension of intra-tumoral diversity, using genomic and epigenomic data (Cell, 2013; Cancer Cell, 2014; Nature, 2015, Nature Communications, 2016, Nature Communications, in press). Notably, we have shown that evolution with therapy is virtually universal, and is anticipated by pretreatment intra-tumoral diversity. To overcome this diversity, effective combination therapy is needed. However, combination therapy design is based on limited data (overlapping toxicities, qualitative pre-clinical data), and fails to engage with the large combinatorial space of drugs, dosages, and schedules. In addition, we lack the tools to robustly identify epigenetic modifications that contribute to cancer evolution with therapy (epidrivers), or study evolution in relation to its micro-environment.
We aim to transform cancer therapy by building the knowledge infrastructure required for combination therapy to directly anticipate and address evolution. To do so, we will measure clonal growth rates directly from serial patient samples collected under the umbrella of phase I/II clinical trials. We will utilize these data to prioritize and optimize combination therapy in a data-driven, iterative, personalized fashion. Furthermore, our studies show that genetic clonal complexity does not fully account for the evolutionary plasticity of CLL. We will apply cutting-edge statistical inference to longitudinally sampled CLLs, profiled with bulk and single-cell bisulfite sequencing. Through these studies, we will identify candidate epigenetic and micro-environmental determinants of evolution in response to therapy. Collectively, this novel set of tools will replace current empiric therapy trial design, which is based on qualitative data and takes years to yield insights. Instead, our foundational work will support data-driven combinatorial therapy. The genetic, epigenetic, and environmental evolutionary potential of CLL will be mapped and personalized combinations will be nominated. Precision, real-time measurements of clonal kinetics will enable us to continuously optimize the drug combinations, dosages, and schedules in order to maximize overall efficacy and overcome the central obstacle of tumor evolution.
As cancer treatments become more effective, the emerging major obstacle to achieving a cure is the ability of malignant cells to adapt to therapy through an intensive evolutionary process. Our proposed project offers to overcome this challenge by providing experimental models and measurements of subpopulation growth rates directly in patients. This information will enable precision customization of anti-cancer drug combinations to prevent disease resistance to therapy.
