Receptor-Driven ERK Pulses Reconfigure MAPK Signaling and Enable Persistence of Drug-Adapted BRAF-Mutant Melanoma Cells

Abstract

Targeted inhibition of oncogenic pathways can be highly effective in halting the rapid growth of tumors but often leads to the emergence of slowly dividing persister cells, which constitute a reservoir for the selection of drug-resistant clones. In BRAF^V600E melanomas, RAF and MEK inhibitors efficiently block oncogenic signaling, but persister cells emerge. Here, we show that persister cells escape drug-induced cell-cycle arrest via brief, sporadic ERK pulses generated by transmembrane receptors and growth factors operating in an autocrine/paracrine manner. Quantitative proteomics and computational modeling show that ERK pulsing is enabled by rewiring of mitogen-activated protein kinase (MAPK) signaling: from an oncogenic BRAF^V600E monomer-driven configuration that is drug sensitive to a receptor-driven configuration that involves Ras-GTP and RAF dimers and is highly resistant to RAF and MEK inhibitors. Altogether, this work shows that pulsatile MAPK activation by factors in the microenvironment generates a persistent population of melanoma cells that rewires MAPK signaling to sustain non-genetic drug resistance.

Keywords: BRAF(V600E) melanoma; MAPK pathway; cancer persistence; kinase inhibitors; kinetic modeling; non-genetic drug resistance; signaling plasticity; systems pharmacology; targeted therapy.

Figure 1.. BRAF^V600E Melanoma Cells Exposed to RAF and MEK Inhibitors Exhibit Spatially Localized Pulsatile ERK Reactivation

(A) pERK levels by immunofluorescence microscopy in A375 cells treated with 1-μM vemurafenib. (B) Single-cell pERK levels in parental (left) and clonal (right) A375 cells treated with vemurafenib for 24 h. (C) Single-cell pERK levels in A375 cells treated with trametinib for 24 h. (D) pERK staining in A375 cells treated with RAF and MEK inhibitors. (E) ERK activity traces (N = 861) quantified by the ERK-KTR reporter from A375 cells treated with vemurafenib (1μM, 24 h). Red numbers correspond to images in (E). (F) Shape of average ERK pulse from 296 live-cell trajectories (solid line indicates mean and shading one standard deviation). (G) Percentage of A375 cells with ERK pulses over a 16-h period. (H) Single-cell ERK activity distributions of A375 treated with DMSO, vemurafenib (1μM) alone, or with cobimetinib (1μM). Colored dots denote ERK activity levels in the colored traces in (E). (I) Percentage of ERK pulsing events that are synchronous or alternate in neighboring cells. (J) Heatmap of ERK activity in cells treated with vemurafenib (1μM, 24 h). Arrows denote ERK pulses; images of these pulses are shown on the right for the ERK:KTR CFP channel (upper row) and corresponding ERK activity (cCFP/nCFP ratio) (lower panel). Lysis of a dead cell preceding a synchronous ERK pulse event is shown in the magnified insert. (K) Heatmap and image snapshots of ERK activity for neighboring cells exhibiting alternate ERK pulsing, shown as in (J).

Figure 2.. ERK Pulses Enable Slow Proliferation of Drug-Adapted BRAF^V600E Melanoma Cells

(A) Heatmap of geminin:RFP intensity for individual cells treated with DMSO, vemurafenib (1μM) alone, or in combination with a saturating dose of cobimetinib (1μM) for 24 h and then imaged for 4 days. (B) Heatmap of geminin:RFP intensity for individual cells that escaped G0/G1 cell-cycle arrest after 4 days of vemurafenib (1μM) treatment. (C) ERK activity and geminin:RFP intensity for two exemplary cells from (B). (D) Percentage of pERK-high cells and EdU positive A375 cells treated with vemurafenib (1μM) and varying cobimetinib concentrations for 4 days. (E) Single-cell pERK distributions for three BRAF^V600E melanoma cell lines treated with 1-μM vemurafenib alone (1^st row) or in combination with 1-μM cobimetinib (2^nd row) and percentage of pERK-high cells (3^rd row) and cell counts (4^th row) during treatment. (F) Schematic depiction of relationship between ERK pulsing and slow cycling in drug-adapted BRAF-mutant melanoma cells.

Figure 3.. Drug-Adapted BRAF^V600E Melanoma Cells Can Use Multiple Receptors to Induce ERK Pulses

(A) pERK distributions of A375 cells treated for 36 h with vemurafenib followed by addition of inhibitory drugs for 2 h. (B) Cell count (1^st panel), single-cell pERK distribution (2^nd panel), and percentage of pERK-high cells (3^rd panel) for A375 cells treated for 4 days with single or combinations of MAPK cascade targeted inhibitors. (C) Schematic of reduction in spontaneous ERK pulses achieved by co-targeting MAPK cascade components in addition to vemurafenib. (D) Receptor abundance by immunofluorescence (top) and ELISA (bottom) in A375 cells with and without vemurafenib treatment for 24 h. (E) pERK in cells exposed to different concentrations of growth factors in A375 cells exposed to vemurafenib for 24 h. (F) pERK levels in cells exposed to growth factors (100 ng/mL) for 15 min at different times after the addition of vemurafenib. Shading interpolates maximal pERK levels. (G) pERK levels for parental, CRISPRa EGFR overexpressing (aEGFR1 and aEGFR2), and CRISPRi EGFR downregulated cell lines and exposed to vemurafenib for 24 h and then to EGF (100 ng/mL). (H) EGFR localization by immunofluorescence microscopy in normal and EGFR-overexpressing cells exposed to vemurafenib for 24 h and then to EGF (100 ng/mL). (I) Population-level and single-cell pERK levels in A375 and A375 EGFR-overexpressing (aEGFR2) cells exposed to vemurafenib (1μM, 24 h) and then to 5 min of EGF over a 10⁶-fold concentration range. (J) EGFR abundance by immunofluorescence microscopy at the membrane of A375 or A375 EGFR-overexpressing (aEGFR2) cells exposed for 24 h to 1-μM vemurafenib. (K) Transcript levels in A375 cells exposed to vemurafenib for 24 h and then EGF (100 ng/mL).

Figure 4.. Quantitation and Effects of Negative Feedback Relief and Receptor Activation on MAPK Signaling in Melanoma Cells Exposed to RAF and MEK Inhibitors

(A) Schematic of the ERK pathway in BRAF^V600E cells; negative feedback regulators are shown in blue. (B) Absolute abundances of MAPK pathway proteins by proteomics following 24-h exposure to vemurafenib at the doses indicated in D. (C) Changes in mRNA levels by RNA-seq for the same proteins. (D) Phosphorylation of key regulatory sites on selected proteins by phosphoproteomics. (E–G) pERK and pMEK levels following 24-h exposure of cells to vemurafenib or dabrafenib followed by addition of EGF (100 ng/mL) for 5 min (E), of NRG1, FGF8, and HGF (100 ng/mL) for 5 min (F), or EGF (100 ng/mL) for indicated amount of time (G). (H–J) pERK and pMEK levels following 24-h exposure of cells to cobimetinib or trametinib followed by addition of EGF (100 ng/mL) for 5 mins (H), or NRG1, FGF8, or HGF (100 ng/mL) for 5 min (I), or EGF (100 ng/mL) for the times indicated (J). (B–D), error bars indicate standard deviations from four replicates. (E–J), error bars indicate standard deviations from two replicates.

Figure 5.. Computational Modeling of ERK Pulsing during Combined RAF and MEK Inhibition

(A) Species and interactions in the MARM1 computational model. (B) Model simulations of protein species (solid lines indicate median values and shades interquartile ranges) and their fits to experimental data (dots). pERK (upper panels) and pMEK (low panels) levels after exposure of cells to EGF for 5 min in the presence of varying concentrations of vemurafenib or cobimetinib for 24 h. (C) Model simulations of pERK levels and protein abundances in cells treated with vemurafenib for 24 h. t = 0 represents the time of EGF addition in the left panel and of vemurafenib in the right panel. (D) Distributions of kinetic parameters from 100 estimation runs. Dot indicates median, thick bars interquartile ranges, and thin lines the minimum and maximum values in log₁₀ scale. Note that the rate for BRAF^V600E phosphorylation of MEK when MEK is bound by MEKi was set to zero. (E and F) Calculated net receptor-driven ERK signaling as the difference between pERK levels in EGF stimulated cells (5 min) and unstimulated cells following pretreatment for 24 h with the indicated RAF (E) or MEK (F) inhibitors. Measurements (colored lines) are shown overlaid to model fits (dotted black line). (G) Model predictions (left panels) and experimental validation (right) of pERK (upper rows) and pMEK (lower rows) in cells treated for 24 h with vemurafenib (1μM) plus cobimetinib with or without subsequent addition of EGF, NRG1, FGF8, or HGF (100 ng/mL for 5 min). For simulations, solid lines indicate median values and shades interquartile ranges. For experiments, error bars indicate standard deviations from two replicates. (H) Model simulation (top panels) and experimental validation (bottom panels) of pERK levels following 24-h exposure to vemurafenib plus cobimetinib in A375 cells exposed to EGF (100 ng/mL for 5 min) (left column), without addition of EGF (central column), and subtraction of the two to estimate net receptor-driven ERK activity (right column).

Figure 6.. ERK Pulsing in Mouse Xenografts and Replication Stress in Drug-Adapted Melanoma Cells

(A) Immunofluorescence microscopy of pERK in A375 mouse xenografts treated either with vehicle alone or dabrafenib plus trametinib for 5 days. (B) Intravital imaging of A375 fluorescent reporter cells in mouse xenografts treated for 48 h with vehicle alone or dabrafenib plus trametinib. For drug-naive tumors, a ERK-KTR:CFP field of view is shown; for tumors in drug-treated animals a field of view and two time courses are shown with ERK-KTR:CFP above and H2B:YFP below. Arrows indicate cells identified to undergo an ERK pulse. (C) Expression of genes involved in DNA repair and of DNA polymerases in A375 cells treated for 24 h with vemurafenib at different doses (line plots, left) or with 1-μM vemurafenib (bar plots, right). (D) EdU and p-γH2AX staining in A375 cells treated for 96 h with vemurafenib (1μM) alone or in combination with saturating doses of cobimetinib (1μM). Percentage of cells having incorporated EdU or positive for γH2AX staining is shown. Error bars indicate standard deviations from three replicates.

Figure 7.. Differential Inhibition of Oncogenic and Physiological MAPK Signaling and Persistence of BRAF-Mutant Melanoma Cells through Spontaneous Receptor-Driven ERK Pulses

(A) Schematic of spontaneous RTK-mediated ERK pulses that promote slow proliferation of BRAF^V600E melanoma cells in which oncogenic MAPK signaling is profoundly inhibited by RAF and/or MEK inhibitors. (B) Presence in BRAF^V600E melanoma cells of two different configurations of the MAPK signaling cascade. The oncogenic configuration inhibits activation of the physiological configuration through negative feedback regulators such as DUSPs, SPRY proteins and phosphorylation of SOS1. When the drugs are present and BRAF^V600E signaling is inhibited, the levels of negative regulators fall, allowing the physiological configuration, which is resistant to both RAF and MEK inhibitors, to transduce growth factor signals operating in a paracrine/autocrine manner.