Skip to contentSkip to filters

We use cookies for analytics.

A Mathematical Model of Evolution of Drug-Induced Resistance

Eduardo Sontag

Northeastern University

Recording

49 min

Abstract

Resistance to chemotherapy is a major impediment to successful cancer treatment. Resistance has classically been thought to arise mainly through random genetic mutations followed by Darwinian expansion of mutated cells, but recent experimental evidence suggests that progression to resistance can instead be induced by the therapeutic agent itself, through genetic changes or through epigenetic alterations that cause otherwise drug-sensitive cells to undergo phenotype switching. This notion further complicates the design of treatment protocols that minimize the risk of evolving resistance.

The talk presents a mathematical modeling framework that incorporates both spontaneous and drug-induced resistance. The model shows that a drug's ability, or inability, to induce resistance can produce qualitatively different responses to the same dose and delivery schedule, and a structural identifiability theorem establishes when the resistance induction rate of a drug can be determined for a given protocol. An associated optimal control problem is characterized precisely as a concatenation of bang-bang and path-constrained arcs using the Pontryagin Maximum Principle and differential Lie algebraic techniques, and a structural identifiability analysis shows that patient-specific parameters can be measured and used to design optimal therapies before treatment begins. A slightly modified model gives excellent fits to time-resolved in vitro data: from total cell counts it recovers the relative proportions of sensitive and resistant subpopulations and quantifies their dynamics as a function of dose, with predictions validated on doses not used for fitting. Combined with optimal control, the model is used to discover dosing strategies that may lead to better outcomes, measured by lower total cell volume.

Topics