| dc.description.abstract | Small cell lung cancer (SCLC) is a phenotypically heterogeneous disease, comprising multiple cellular subtypes within a tumor that exhibit differential sensitivity to drug treatments. SCLC heterogeneity is hypothesized to be responsible for the dismal 7% five-year survival rate for this disease. Phenotypic plasticity, a recognized phenomenon, enables populations of cells to withstand treatment entirely or to cycle in a drug-tolerant state long enough to acquire resistance mutations. Tumors reaching a new population equilibrium or re-attaining equilibrium after a disturbance is the likely reason for many patients’ tumor recurrence. A cellular population dynamics model of SCLC will enable prediction of tumor growth and treatment response based on subpopulation proportions, providing the opportunity to design novel experiments or test new treatment strategies. However, SCLC behavior differs depending on the biological context in which it is studied, making design of a single population dynamics model difficult. Thus, we use Bayesian multimodel inference to generate and evaluate thousands of SCLC mechanistic models, each with subsets of the SCLC phenotypic behaviors seen in varying biological contexts. We explore mechanistic hypotheses of SCLC behavior, evaluating heterogeneity, lineage plasticity, and cell-cell interactions. Our approach directly enables estimation of how informative the data are for a given hypothesis, and whether multiple hypotheses support the data equally well. The resulting probabilistic view of subtype behaviors indicates a high likelihood of phenotypic transitions across the differentiation hierarchy in SCLC. We perform additional analyses on phenotypic changes during tumor growth and treatment using time-course mass cytometry data in treated and untreated patient-derived xenograft (PDX) models. This data demonstrates the existence and outgrowth of a small stem-like cell subpopulation in SCLC PDXs, likely contributing to treatment resistance. The combination of an increased probabilistic understanding of SCLC growth mechanisms, and the demonstration of phenotypic transitions in an independent, time-course data series, highlights what mechanistic knowledge is supported by data and which desired knowledge requires further experiments. With results predicting that any SCLC subtype can repopulate the tumor post-treatment, modulation of tumor composition and effects on treatment resistance can be tested experimentally to increase knowledge and decrease the burden of this aggressive disease. | |
