Many theories have been proposed describing the single-cell dynamics of chemotherapy response and expansion of resistant clones. These usually invoke the presence of low frequency somatic mutations or the de novo acquisition of new mutations. In contrast to these predominantly genetic mechanisms, we have now utilised mathematical modelling and longitudinal single-cell imaging to demonstrate that a chemoresistant population of cancer cells can emerge solely through the inherently noisy process of gene expression, which is amplified by the non-linear behaviour of apoptotic signalling pathways.
High-risk neuroblastoma is an aggressive, highly chemoresistant childhood tumour. We previously demonstrated that in silico, patient-specific modelling of apoptotic signalling could stratify neuroblastoma patient cohorts and provide robust biomarkers of patient survival. We now show that application of this patient-level model to single-cell populations also predicts the presence of this innately chemoresistant cell population, which cannot activate sufficient drug-induced signalling to reach an in-built apoptotic threshold.
Using JNK activity biosensors with longitudinal high-content and intravital imaging, we have now confirmed that this stochastic population of chemoresistant cells exist prior to treatment and are selected for during chemotherapy treatment. Furthermore, using matched PDX models established at diagnosis and relapse from the same patients, we demonstrate that a memory of this resistant state is maintained through epigenetic remodelling. Consequently, we show that priming neuroblastomas with an HDAC inhibitor cannot erase the memory of this resistant state within relapsed neuroblastomas, but improves response in the first-line setting by restoring drug-induced JNK activity within the chemoresistant population of treatment naïve tumours.