Cyclic cell state transition is associated with the adaptive resistance to BRAF inhibition in melanomas

Abstract

Cancers commonly develop resistance against chemotherapeutics or targeted therapies through either genetic mechanism or adaptive responses that commonly involve epigenetic reprogramming and activation of compensatory signaling pathways. Recent examples of such an adaptive response involve the treatment of BRAF mutant melanoma cancer patients with inhibitors of the BRAF/MAPK pathway. Several studies attribute it to a reversible cell state transition from a proliferative melanocytic state towards invasive neural crest-like and mesenchymal-like states. The capacity of certain BRAF mutant melanoma patient tumors, and the cell lines derived from them, to adapt to targeted therapies bears some conceptual similarities to adaptive responses obsd. in other cancers. Further, the response of the cells that adapt to BRAF inhibition is highly variable depending on phenotypic plasticity of the cells, the length of drug exposure, and the dosing strategy. As such, general physiochem. approaches that provide quant., predictive models of such transitions may provide a common conceptual framework for understanding the large genome-scale gene expression level changes. Through analyzing the kinetic trajectory of the reversible, adaptive response of a highly plastic, patient-derived BRAF^(V600) mutant melanoma cell line to BRAFi, using genome-wide expression of these cells at 15 time points over a period that included 29 days of drug treatment, followed by either an addnl. 30 days of treatment, or 35 days of drug release, we found that the cells adapt to the drug treatment in two stages, terminating in a mesenchymal-like state after almost a month of treatment. At that stage, continued drug exposure induced no further changes in cell phenotype. Following drug release, the cells reverted back to the original, drug naive phenotype, but the reverse pathway took a different trajectory than the forward pathway. Thermodynamically-motivated Surprisal Anal. was utilized to identify two constraints of this cyclic cell state transition to construct a free energy surface. A kinetic ODE model of the two unbalanced processes revealed that the process assocd. with the long-term adaptive transition was driven the other process reflective of a short-term drug stress response. Such kinetic model further allowed identifying driving transcription factors and epigenetic regulations that prompt the cyclic cell state transition towards adaptive resistance.