Deconstructing the Role of the Microenvironment on Drug Efficacy in a Brain-Mimetic Platform for Cutaneous Metastatic Melanoma

Abstract

Although survival in patients with malignant melanoma has significantly improved due to therapeutic interventions based on the molecular basis of tumor etiology, durable responses in the face of metastatic disease are rarely realized. A “systems pharmacological” approach to uncover drug potency at the physically distinct stages of the metastatic cascade is required. We modeled disparate microenvironments in the brain; the perivascular niche and hyaluronic acid (HA) rich parenchyma, to assess contextual drug efficacy. These two microenvironments are not only differ in composition, but in dimensionality, with the perivascular niche inducing a 2D morphology in cells, while the HA-rich parenchyma leads to 3D cellular clusters. These in vitro models recapitulated in vivo morphology and motility for an isogenic, human model of melanoma metastatic progression. By independently modulating adhesion strength and ECM composition, we found that ERK inhibition decreased cell adhesion, whereas BRAF inhibition was only effective when combined with an ERK inhibitor. BRAF and ERK inhibition individually reduced cell motility in the less metastatic clone, with a lesser effect on the more metastatic clone. We observed that cells are resistant to BRAFV600E inhibition when cultured in 3D Fibronectin rich HA hydrogels, but Laminin rich HA-gels offered no protection. The opposite held true for ERK inhibition. These data reinforce that a dynamic microenvironment not only contributes to systemic metastasis, but also significantly modifies drug efficacy.