Exploiting Drug Addiction Mechanisms to Select against MAPKi-Resistant Melanoma

Abstract

Melanoma resistant to MAPK inhibitors (MAPKi) displays loss of fitness upon experimental MAPKi withdrawal and, clinically, may be resensitized to MAPKi therapy after a drug holiday. Here, we uncovered and therapeutically exploited the mechanisms of MAPKi addiction in MAPKi-resistant BRAF^MUT or NRAS^MUT melanoma. MAPKi-addiction phenotypes evident upon drug withdrawal spanned transient cell-cycle slowdown to cell-death responses, the latter of which required a robust phosphorylated ERK (pERK) rebound. Generally, drug withdrawal-induced pERK rebound upregulated p38-FRA1-JUNB-CDKN1A and downregulated proliferation, but only a robust pERK rebound resulted in DNA damage and parthanatos-related cell death. Importantly, pharmacologically impairing DNA damage repair during MAPKi withdrawal augmented MAPKi addiction across the board by converting a cell-cycle deceleration to a caspase-dependent cell-death response or by furthering parthanatos-related cell death. Specifically in MEKi-resistant NRAS^MUT or atypical BRAF^MUT melanoma, treatment with a type I RAF inhibitor intensified pERK rebound elicited by MEKi withdrawal, thereby promoting a cell death-predominant MAPKi-addiction phenotype. Thus, MAPKi discontinuation upon disease progression should be coupled with specific strategies that augment MAPKi addiction.Significance: Discontinuing targeted therapy may select against drug-resistant tumor clones, but drug-addiction mechanisms are ill-defined. Using melanoma resistant to but withdrawn from MAPKi, we defined a synthetic lethality between supraphysiologic levels of pERK and DNA damage. Actively promoting this synthetic lethality could rationalize sequential/rotational regimens that address evolving vulnerabilities. Cancer Discov; 8(1); 74-93. ©2017 AACR.See related commentary by Stern, p. 20This article is highlighted in the In This Issue feature, p. 1.

Figure 1

MAPKi-resistant melanoma display distinct drug-addiction phenotypes characterized by slow-cycling versus cell-death responses. (A) Clonogenic growth of double-drug resistant or DDR (^MUTBRAF) or single-drug resistant or SDR (^MUTNRAS) melanoma cell lines plated 24 hrs with BRAFi (vemurafenib)+MEKi (selumetinib) at 1 µM or MEKi (trametinib) at 0.1 µM followed by 7 d with (On) or 7 and 18 d without (Off) inhibitor withdrawal. (B) Temporal vital images of MAPKi-resistant or R-lines on or off BRAFi+MEKi. (C–E) Percentages of annexin V/PI-positive dead cells (C), CFSE dye dilution patterns (D), and levels of SA-βgal staining (E) in R-lines on or off MAPKi(s) for 6 d. Loading control (D) refers to the intensity of the CFSE dye initially loaded into the cells. (F) Correlation between fold-changes (FC) in CFSE dye dilution and % cell death off vs. on MAPKi(s).

Figure 2

Excessive ERK activation induces DNA damage and AIF-mediated death in the cell-death predominant MAPKi-addiction phenotype. (A) Heatmap showing the gene set variance analysis scores of differentially enriched gene signatures (median log fold-change (FC) ≥ 1.25 in the off-drug condition compared to the on-drug condition in either cell line; additionally, differential enrichment between the cell-death predominant and slow-cycling predominant R-lines must also be higher (by at least 25%) in the off-drug condition than in the on-drug condition). (B–C) Levels of AIF, PARP, and p-H2AX measured by Western blots (WB) (B) or immunofluorescence (IF) (C) in R-lines on or off MAPKi(s) for 1 and 3 d or 3 d, respectively. For b, mitochondrial (Mito), nuclear (Nucl), cytoplasmic (Cyto) cellular fractions or whole cell lysates (WCL). For c, nuclei visualized by DAPI; ruler, 20 µm. (D) DNA strand break measurements by the comet assay of R-lines on or off MAPKi(s) for 3 d. Tail length/moment FCs were quantified (n=5; mean ± SDs). (E) Temporal levels of p-ERK, p-H2AX and PARP in the nuclear fraction of R-cell lysates measured by WBs at indicate hrs off MAPKi(s). (F–G) Levels of indicated proteins by WBs in the nuclear (F, G) or WCL (F) fractions of R-lines, on and off MAPKi(s), transduced with control (V) or AIF (F) or H2AX (G) shRNA lentiviruses. (H–M) Percentages of annexin V/PI-positive dead cells (H, I), CFSE dye dilution patterns (J, K), and viable cell counts (L, M) in R-lines on or off MAPKi(s) for 6 d, transduced with control (Vector) or AIF (J, L) or H2AX (K, M) shRNA lentiviruses. For l, M, n=6; mean ± SDs; ***p <0.001 based on ANOVA. (N–O) Sub-cellular localization of PAR (N) or levels of PARG (O) by IF in R-lines. Nuclei visualized by DAPI; ruler, 20 µm. (P) PARG and p-ERK levels by WBs in the indicated fractions of M249 DDR5 or nuclear fractions of indicated additional R-lines. For WB, GAPDH, loading control.

Figure 3

Impairing DNA damage repair augments MAPKi-addiction. (A) Western blot (WB) analysis of p-H2AX levels in slow-cycling predominant R-lines on or off MAPKi(s), with or without ATMi and/or PARPi treatment for 3 d. (B–C) Clonogenic growth (B) or percentages of annexin V/PI-positive dead cells (C) in slow-cycling predominant R-lines on or off MAPKi(s) for 5 d, with or without ATMi and/or PARPi treatment. (D) WB analysis of BRCA1 and p-H2AX levels in in slow-cycling predominant R-lines, transduced with control (V) or BRCA1-specific shRNA lentiviruses, on or off MAPKi(s), with or without PARPi treatment for 3 d. (E–F) R-line cells from D were subjected to the same assays as in B and C, respectively. For WBs, TUBULIN, loading control. (G) Levels of indicated proteins by WBs in mitochondrial (MITO), nuclear (NUCL) subcellular fractions or whole-cell lysates (WCL) of slow-cycling versus cell-death predominant R-lines, on and off MAPKi(s), with or without ATMi and PARPi treatment for 3 d. HSP60, mitochondrial fraction control; Histone H3, nuclear fraction control. (H–I) Clonogenic growth (H) or percentages of annexin V/PI-positive dead cells (I) in cell-death predominant R-lines on or off MAPKi(s) for 6 d and 16 (only in H), with or without ATMi and/or PARPi treatment. (J–L) Levels of caspase-3 activity (J, K) or clonogenic growth (L) with indicated MAPKi(s) treatment, with or without ATMi and PARPi, with or without caspase-3 inhibitor, Z-DEVD-FMK (20 µM), in slow-cycling predominant R-lines (J) for 3, 6 d or cell-death predominant R-lines (K) for 3 d. For J, K, n=5; mean ± SDs; ***p <0.001 based on ANOVA. Staurosporine (1 µM) used to induce caspase-3 activity and death.

Figure 4

Paradoxical ERK activation by BRAFi potentiates drug addiction in MEKi-resistant ^MUTNRAS melanoma. (A–B) Western blot levels of p-ERK, ERK, and loading control TUBULIN in ^MUTNRAS parental and isogenic MEKi-resistant SDR-lines with indicated hrs on MEKi/trametinib (0.1 µM) treatment (A) or in SDR-lines with indicated hrs off MEKi/trametinib (0.1 µM) treatment, with or without BRAFi/vemurafenib (5 µM) treatment (B). (C–D) Clonogenic growth (C) and percentages of annexin V/PI-positive dead cells (D) in ^MUTNRAS SDR-lines on or off MEKi/trametinib (0.1 µM) for 6 d, with or without BRAFi/vemurafenib (5 µM) Inset for M207 SDR1 off MEKi for 16 d. (E) Levels and/or sub-cellular localization of p-H2AX, PAR, and AIF in cell-death predominant MEKi-addicted R-lines on or off MEKi for 3 d, with or without BRAFi (5 µM) treatment.

Figure 5

Excessive ERK and DNA damage induce regression of MEKi-resistant PDX tumors. (A) Tumor volumes (mean ± SEM) of a ^MUTNRAS PDX transplanted in NSG mice in response to daily gavage with the vehicle (n = 3) or trametinib (5 mg/kg; n = 5). Asterisk indicates the resistant tumor, R1, which was serially transplanted for experiment in B. (B) R1 trametinib-resistant ^MUTNRAS PDX tumors were grown in NSG mice for 55 days with daily trametinib (5 mg/kg) gavage until initiating indicated daily treatments or regimens (n = 4 in each group). Tumor volumes are shown as means ± SEM. P values, Student’s t-test. MEKi, trametinib (5 mg/kg); BRAFi, vemurafenib (100 mg/kg). (C) Pictures of tumors from three experimental groups in B at day 63. (D) Tumor weights (means ± SEM; p value, unpaired t-test) of three experimental groups in B. (E) Levels of indicated proteins in representative tissue sections of tumors in C. Ruler, 20 µm. (F–J) As in A to E except experiments used a distinct PDX model harboring ^S365LBRAF and individual tumor growth curves were plotted separately in F. The MEKi-resistant tumor (R3) which arose first was fragmented, serially passaged and used in G. (K–L) R1 trametinib-resistant ^MUTNRAS PDX tumors were serially passaged in NSG mice with daily trametinib (5 mg/kg) gavage until segregation into seven groups (1, 2, 3, 5, n=3 per group; 4, 6, 7, n=5 per group). Tumor volumes and weights are shown as means ± SEM. P values, Student’s t-test.

Figure 6

BRAF and PARP inhibitors augment MEKi addiction of ^MUTNras murine melanoma in an immune competent host. (A) Tumor volumes (mean ± SEM) of TpLN^61R murine melanoma cells transplanted in C57BL/6 mice in response to daily gavage with the vehicle (n = 6) or trametinib (5 mg/kg; n = 6). One resistant tumor on day 36 was dissociated and cultured as a MEKi-resistant cell line (NILR2R). (B) Western blot levels of p-ERK, ERK, and the loading control GAPDH in the MEKi-resistant ^MUTNras SDR line, NILR2R, with indicated hrs off MEKi/trametinib (0.1 µM) treatment, with or without BRAFi/vemurafenib (1 µM) treatment. (C) Analysis of NILR2R protein lysates by Western blots of p-ERK and p-H2AX levels on or off MAPKi(s), with or without BRAFi (1µM) and/or PARPi (0.2µM) treatment for 3 d. (D–E) Clonogenic growth (D; 8 d) and percentages of Annexin V/PI-positive dead cells (E; 5 d) in NILR2R, on or off MEKi/trametinib (0.1 µM), with or without BRAFi/vemurafenib (1 µM) and/or PARPi (0.2µM) treatments. For D, cultures were seeded at 30K cells per well, except for the third row where cultures were seeded at 150K cells per well. (F) Levels and/or subcellular localization of p-H2AX, PAR, and AIF in NILR2R, on or off MEKi for 3 d, with or without BRAFi and/or PARPi treatment. Ruler, 20 µm. (G) Trametinib-resistant NILR2R cells were transplanted subcutaneously in C57BL/6 mice with daily trametinib (5 mg/kg) gavage until segregation into five groups (n=6 per group). Tumor volumes are shown as means ± SEM. P values, Student’s t-test. (H) Pictures of tumors from the first four experimental groups (mice sacrificed due to tumor ulceration) in G at day 26. (I) Tumor weights (means ± SEM; p value, unpaired two-way t-test) of the first four experimental groups in G.

Figure 7

Strategies to select against MAPKi-resistant melanoma. Schematic showing MAPKi-addiction phenotypes being driven by p-ERK rebound levels and potential therapeutic strategies (enhancing p-ERK or impairing DNA damage repair) that promote tumor cell-death (apoptosis or parthanatos) or regression.