Genetic mutations that cause drug failure are a major obstacle in many diseases including cancer, HIV, cytomegalovirus, and tuberculosis. Therapy-induced selection for resistance-conferring aberrations can arise either from new mutations or from those that pre-existed in a small subpopulation of the cells or viruses. This latter idea of pre-existing subclonal drug resistance is highly understudied across diseases, and there is no general clinical consensus on the best way to implement counter-resistance therapies. A large part of this is due to technical difficulties in detecting the subpopulations at both sufficient resolution and high enough throughput. Here, we leverage a novel high-sensitivity DNA mutation detection technology, multiplex blocker displacement amplification, and melanoma as a model system to study subclonal drug resistance for multiple genes inhundreds of pre-therapy patient samples. We will pair this clinical study with novel mouse models of subclonal resistance to optimize risk-reward strategies for counter-resistance therapies. Our preliminary data from melanoma patients are consistent with data from other cancers suggesting that very low allelic-frequency subclonal resistance mutations could pre-exist in over a third of patients' tumors. Therefore, our overall approach is aimed at determining how to best treat patients with potential subclonal resistance, first by improving mutation detection in patients and second by determining which mutation-positive patients would most benefit from optimally-timed counter-resistance interventions. Although we start with melanoma as a model system, our approach will serve as a broadly-applicable blueprint for recognizing and overcoming pre-existing subclonal resistance.
Small subpopulations of diseased cells may already harbor drug-resistance mutations, eventually causing therapy failure. How often such cells exist across different types of disease is largely unknown, and no consensus yet exists on how to optimally treat them. In this proposal, we will use melanoma as a model system, coupled with a novel high-sensitivity, high-throughput mutation detection technology, to identify how often pre-existing drug resistance mutations occur in untreated patients and determine their association with therapy outcomes; in parallel we will use mouse models to understand the behavior of these pre-existing cells and determine the best way to prevent them from causing therapy failure. Our overall approach will be readily applicable to other diseases that show drug failure due to genetic mutations, and our results will form a key proof-of-principle for a new paradigm of treating with upfront counter-resistance therapies.
