Tumor cell plasticity and heterogeneity represent major challenges to achieving effective treatment in cancers. In the context of glioblastoma (GBM), the most prominent and lethal primary brain tumor in adults, tumor cell plasticity and phenotypic shifts are driven by epigenetic mechanisms that can lead to the emergence of resistance, i.e., acquired resistance, in a subpopulation of tumor cells to a drug to which those tumor cells were originally sensitive. We have previously observed drug-induced phenotypic changes in GBM stem-like cells (GSCs), a rare subpopulation of tumorigenic cells that undergo a proneural-to-mesenchymal transition (PMT) as a mode of therapy evasion. Here, we examined the dynamics of PMT in two patient-derived GSC populations (PD-GSCs), one sensitive to and another resistant to pitavastatin, a drug that has anti-proliferative effects on glioma cells. We applied a variety of bioinformatic analyses and optimal transport analysis on time-course single-cell RNA-seq profiles of PD-GSC drug responses to identify trajectories along which cells transition into a drug-resistant state. Our results revealed that sensitive PD-GSCs traversed multiple trajectories, along which cells underwent PMT, ultimately leading to a homogenization of the surviving cell population and development of sustained resistance. In contrast, non-responsive PD-GSCs, which exhibited an incoherent population structure with no significant homogenization or emergence of a predominant trajectory along which phenotypic transitions would have occurred. The current work applies a systems biology approach to evaluate the causative mechanisms driving GSC dynamics that lead to acquired resistance.
