The MEK5/ERK5 pathway promotes the activation of the Hedgehog/GLI signaling in melanoma cells

Abstract

Purpose: Malignant melanoma is the deadliest skin cancer, with a poor prognosis in advanced stages. We reported that both Hedgehog-GLI (HH/GLI) and Mitogen-activated protein Kinase (MAPK) extracellular signal-regulated kinase 5 (ERK5) pathways promote melanoma growth, and that ERK5 activation is required for HH/GLI-dependent melanoma cell proliferation. Here, we explored whether ERK5 regulates HH/GLI signaling.

Methods: Both genetic (using ERK5-specific shRNA) and pharmacologic (using the ERK5 inhibitors JWG-071 and AX15836, and the MAPK/ERK kinase 5, MEK5 inhibitors GW284543 and BIX02189) targeting approaches were used. Luciferase assay using the GLI-binding site luciferase reporter was performed to evaluate GLI transcriptional activity. A constitutively active form of MEK5 (MEK5DD) was used to induce ERK5 activation. 3D spheroid assays were performed in melanoma cells.

Results: Genetic and pharmacologic ERK5 inhibition reduces GLI1 and GLI2 protein levels and transcriptional activity of endogenous HH/GLI pathway induced by the agonist SAG in NIH/3T3 cells. In these cells, MEK5DD overexpression potentiates transcriptional activity of endogenous HH/GLI pathway induced by SAG, whereas ERK5 silencing prevents this effect. Consistently, MEK5DD overexpression increases GLI1 and GLI2 protein levels. In melanoma cells, ERK5 silencing reduces GLI1 and GLI2 mRNA and protein levels and inhibits GLI transcriptional activity. MEK5DD further increases the transcriptional activity of the HH/GLI pathway and GLI1 protein levels. Combination of GLI and MEK5 inhibitors is more effective than single treatments in reducing melanoma spheroid growth.

Conclusions: MEK5-ERK5 is an activator of GLI transcription factors, and combined targeting of these pathways warrants further preclinical investigation as a potential innovative therapeutic approach for melanoma.

Keywords: ERK5/MAPK7; GLI; Hedgehog; Melanoma; Targeted therapy.

Fig. 1

ERK5 silencing inhibits endogenous HH/GLI pathway in NIH/3T3 cells. (A-B) Cells were transduced with lentiviral vectors carrying control non-targeting shRNA (shNT) or murine ERK5-specific shRNA (shERK5-1), and then serum-deprived for 24 h (A) or treated with vehicle (-) or SAG for 48 h (B) before lysis and Western Blot. Graphs show averages of relative integrated density (RID) ± SD (n = 3). (C) Cells transduced as above (shNT, shERK5-1, shERK5-2) were serum-deprived for 24 h, transfected with a GLI responsive luciferase reporter (GLI-BS) for 12 h and treated with vehicle (-) or SAG for 48 h. Relative luciferase activity (RLU) was firefly/Renilla ratios normalized for control. Data are presented as means ± SD (n = 3)

Fig. 2

ERK5 inhibition reduces the activity of endogenous HH/GLI pathway in NIH/3T3 cells. (A) 24-hour serum-deprived cells were treated with DMSO (Vehicle), JWG-071 (JWG) or AX15836 (AX) at the indicated concentrations for 1 h before exposure to EGF for 10 min. Cell lysates were analysed by Western Blot. The arrow indicates the slower migrating band of phosphorylated ERK5. (B) 24-hour serum-deprived cells were treated with SAG and/or JWG-071 (JWG) for 48 h. Cell lysates were analysed by Western Blot. Graphs show averages of relative integrated density (RID) ± SD (n = 3). (C) 24-hour serum-deprived cells were transfected with a GLI responsive luciferase reporter (GLI-BS) for 12 h and then subjected for 48 h to the indicated treatments. Relative luciferase activity (RLU) was firefly/Renilla ratios normalized for control ± SD (n = 3)

Fig. 3

ERK5 activation increases GLI1/GLI2 protein levels and transcriptional activity in NIH/3T3 cells. (A) Cells transduced with lentiviral vectors carrying control non-targeting shRNA (shNT) or murine ERK5-specific shRNA (shERK5-1) were transfected with pcDNA3.1 (-) or pCMV5-MEK5DD-HA and with GLI reporter (GLI-BS) for 12 h, before being treated with vehicle (-) or SAG for 48 h. Relative luciferase activity (RLU) was firefly/Renilla ratios normalized for control ± SD (n = 3). (B-C) Cells transfected with pcDNA3.1 or pCMV5-MEK5DD-HA were treated with vehicle (-) or SAG for 48 h. Western Blot was performed. Graphs show averages of relative integrated density (RID) ± SD (n = 3). (D) 24-hour serum-deprived cells were treated with SAG and/or BIX02189. Western Blot was performed. Graphs show averages of RID ± SD (n = 3)

Fig. 4

ERK5 silencing reduces mRNA and protein expression levels of GLI1 and GLI2 transcription factors in melanoma cells. (A-B) A375 (A) and SK-Mel-5 (B) cells were transduced with lentiviral vectors carrying control non-targeting shRNA (shNT) or human ERK5-specific shRNA (shERK5-1 or shERK5-2). Five days after infection, cells were lysed and GLI1 and GLI2 mRNA levels determined by Q-PCR. Data are presented as means ± SD (n = 3). (C-D) A375 (C) and SK-Mel-5 (D) transduced with lentiviral vectors carrying shNT or sh-hERK5-1 were lysed and Western Blot was performed. (E-F) Cells transduced as above were lysed five days after infection and PTCH1 and HIP1 mRNA levels were determined by Q-PCR. Data are presented as means ± SD (n = 3)

Fig. 5

MEK5/ERK5 positively regulates the transcriptional activity of endogenous HH/GLI in melanoma cells. (A) Cells transduced with lentiviral vectors carrying control non-targeting shRNA (shNT) or human ERK5-specific shRNA (shERK5-1/shERK5-2) were transfected with pCS2 + MT (-) or pCS2 + MT-GLI1 and GLI-BS for 12 hours. Relative luciferase activity (RLU) was firefly/Renilla ratios normalized for control ± SD (n = 3). (B) Cells transfected as above were treated with vehicle (0) or XMD8-92 for 24 hours. RLU was firefly/Renilla ratios normalized for control ± SD (n = 3). (C) Cells were transfected with pcDNA3.1 (control) or pCMV5-MEK5DD-HA, lysed after 48 hours, and Western Blot was performed on total (Whole), cytoplasmic (Cyt) or nuclear (Nuc) extracts with the indicated antibodies. Graph shows quantification of nuclear GLI1. (D) HeLa cells were transfected with pcDNA3.1 (control) or pCMV5-MEK5DD-HA for 24 hours. Immunofluorescence analysis was performed by staining for HA (HA-MEK5DD, green) and GLI1 (red). Confocal images were analysed to quantify GLI1 nuclear staining, represented in the graph as nuclear Pearson’ Coefficient ± SD (n = 3). (E) Cells were transfected with pcDNA3.1 (control) or pCMV5-MEK5DD-HA and with a GLI-BS for 12 h. RLU was firefly/Renilla ratios normalized for control ± SD (n = 3). (F) Cells transfected as above were lysed after 48 h and GLI1, MEF2C and MEF2D mRNA levels were determined by Q-PCR. Data are presented as means ± SD (n = 3)

Fig. 6

Combined targeting of MEK5 and GLI synergistically reduces the volume of melanoma spheroids. (A-B) A375 (A) and SSM2c (B) spheroids were treated with DMSO (-), GANT-61, GW284543 or BIX02189 or with the combinations at the indicated concentrations. Graphs show the quantification of spheroid volume after 5 days normalized for the time point 0. Data represent mean ± SD from three independent experiments. Representative images of spheroids taken at day 5 are shown. § Bliss independence score (> 0) indicates synergistic effects over single treatments. Bliss score = 0.027 (GANT + GW284543) or 0.008 (GANT + BIX02189) in A375 cells; Bliss score = 0.291 (GANT + GW284543) or 0.327 (GANT + BIX02189) in SSM2c cells