Advertisement

Regulation of Melanoma Progression through the TCF4/miR-125b/NEDD9 Cascade

  • Florian Rambow
    Affiliations
    Institut Curie, PSL Research University, INSERM U1021, Normal and Pathological Development of Melanocytes, Orsay, France

    Université Paris-Sud, Université Paris-Saclay, CNRS UMR 3347, Orsay, France

    Equipe Labellisée Ligue Contre le Cancer, Orsay, France
    Search for articles by this author
  • Audrey Bechadergue
    Affiliations
    Institut Curie, PSL Research University, INSERM U1021, Normal and Pathological Development of Melanocytes, Orsay, France

    Université Paris-Sud, Université Paris-Saclay, CNRS UMR 3347, Orsay, France

    Equipe Labellisée Ligue Contre le Cancer, Orsay, France
    Search for articles by this author
  • Flavie Luciani
    Affiliations
    Laboratory for Molecular Cancer Biology, Center for Human Genetics, University of Leuven, 3000 Leuven, Belgium

    VIB Center for the Biology of Disease, 3000 Leuven, Belgium
    Search for articles by this author
  • Gwendoline Gros
    Affiliations
    Institut Curie, PSL Research University, INSERM U1021, Normal and Pathological Development of Melanocytes, Orsay, France

    Université Paris-Sud, Université Paris-Saclay, CNRS UMR 3347, Orsay, France

    Equipe Labellisée Ligue Contre le Cancer, Orsay, France
    Search for articles by this author
  • Melanie Domingues
    Affiliations
    Institut Curie, PSL Research University, INSERM U1021, Normal and Pathological Development of Melanocytes, Orsay, France

    Université Paris-Sud, Université Paris-Saclay, CNRS UMR 3347, Orsay, France

    Equipe Labellisée Ligue Contre le Cancer, Orsay, France
    Search for articles by this author
  • Jacky Bonaventure
    Affiliations
    Institut Curie, PSL Research University, INSERM U1021, Normal and Pathological Development of Melanocytes, Orsay, France

    Université Paris-Sud, Université Paris-Saclay, CNRS UMR 3347, Orsay, France

    Equipe Labellisée Ligue Contre le Cancer, Orsay, France
    Search for articles by this author
  • Guillaume Meurice
    Affiliations
    Plateforme de Bioinformatique, UMS AMMICA, Gustave-Roussy, Villejuif, France
    Search for articles by this author
  • Jean-Christophe Marine
    Affiliations
    Laboratory for Molecular Cancer Biology, Center for Human Genetics, University of Leuven, 3000 Leuven, Belgium

    VIB Center for the Biology of Disease, 3000 Leuven, Belgium
    Search for articles by this author
  • Lionel Larue
    Correspondence
    Correspondence: Lionel Larue, Normal and Pathological Development of Melanocytes, CNRS UMR3347, INSERM U1021, Institut Curie Bat 110, Centre Universitaire, 91405 Orsay, France.
    Affiliations
    Institut Curie, PSL Research University, INSERM U1021, Normal and Pathological Development of Melanocytes, Orsay, France

    Université Paris-Sud, Université Paris-Saclay, CNRS UMR 3347, Orsay, France

    Equipe Labellisée Ligue Contre le Cancer, Orsay, France
    Search for articles by this author
Open ArchivePublished:March 08, 2016DOI:https://doi.org/10.1016/j.jid.2016.02.803
      Melanoma progression from a primary lesion to a distant metastasis is a complex process associated with genetic alterations, epigenetic modifications, and phenotypic switches. Elucidation of these phenomena may indicate how to interfere with this fatal disease. The role of microRNAs as key negative regulators of gene expression, controlling all cellular processes including cell migration and invasion, is now being recognized. Here, we used in silico analysis of microRNA expression profiles of primary and metastatic melanomas and functional experiments to show that microRNA-125b (miR-125b) is a determinant candidate of melanoma progression: (i) miR-125b is more strongly expressed in aggressive metastatic than primary melanomas, (ii) there is an inverse correlation between the amount of miR-125b and overall patient survival, (iii) invasion/migration potentials in vitro are inversely correlated with the amount of miR-125b in a series of human melanoma cell lines, and (iv) inhibition of miR-125b reduces migratory and invasive potentials without affecting cell proliferation in vitro. Furthermore, we show that neural precursor cell expressed developmentally down-regulated protein 9 (i.e., NEDD9) is a direct target of miR-125b and is involved in modulating melanoma cell migration and invasion. Also, transcription factor 4, associated with epithelial-mesenchymal transition and invasion, induces the transcription of miR-125b-1. In conclusion, the transcription factor 4/miR-125b/NEDD9 cascade promotes melanoma cell migration/invasion.

      Abbreviations:

      miR (microRNA), miRNA (microRNA), MITF (microphthalmia-associated transcription factor), NEDD9 (neural precursor cell expressed developmentally down-regulated protein 9), NHEM (normal human epidermal melanocytes), TCF4 (transcription factor 4), TCGA (The Cancer Genome Atlas)

      Introduction

      Melanoma is an aggressive tumor arising from transformed melanocytes and is responsible for 80% of skin cancer mortalities. More recently, great improvement was observed with the combination of targeted therapy using BRAF and mitogen-activated protein kinase/extracellular signal-regulated kinase inhibitors and immunotherapy using programmed cell death protein 1 (PD1)/programmed death-ligand 1 (PDL1) and cytotoxic T lymphocyte-associated protein 4 (CTLA4) antibodies, respectively (
      • Ugurel S.
      • Rohmel J.
      • Ascierto P.A.
      • Flaherty K.T.
      • Grob J.J.
      • Hauschild A.
      • et al.
      Survival of patients with advanced metastatic melanoma: the impact of novel therapies.
      ,
      • Queirolo P.
      • Picasso V.
      • Spagnolo F.
      Combined BRAF and MEK inhibition for the treatment of BRAF-mutated metastatic melanoma.
      ). The aggressiveness of melanoma is illustrated by its propensity to metastasize and acquire resistance. Melanoma comprises heterogeneous subpopulations of cancer cells with different biological properties. Transcriptomic profiling of melanoma cell lines and melanomas has identified very divergent programs. In vitro, transcriptomic profiles are consistent with a proliferative and invasive melanoma state (
      • Hoek K.S.
      • Eichhoff O.M.
      • Schlegel N.C.
      • Dobbeling U.
      • Kobert N.
      • Schaerer L.
      • et al.
      In vivo switching of human melanoma cells between proliferative and invasive states.
      ). In vivo, a third state, termed immune, was recently identified: it is characterized by an immune-related signature indicating the presence of tumor-infiltrating lymphocytes (
      • Verfaillie A.
      • Imrichova H.
      • Atak Z.K.
      • Dewaele M.
      • Rambow F.
      • Hulselmans G.
      • et al.
      Decoding the regulatory landscape of melanoma reveals TEADS as regulators of the invasive cell state.
      ). The latest transcriptomic classification of melanoma samples describes the same three subgroups as microphthalmia-associated transcription factor (MITF)-low (invasive), keratin (proliferative), and immune (immune) (
      Cancer Genome Atlas Network
      Genomic classification of cutaneous melanoma.
      ). A recent melanoma progression model takes these different states into account and suggests that melanoma cells can reversibly switch between the corresponding transcriptional programs upon cues from the microenvironment; this phenomenon contributes to intratumor heterogeneity. The phenotype switching model is gaining ground as an alternative genetics/epigenetics mechanism of resistance. MITF-low (invasive) melanoma cells are more resistant than the others to BRAF, mitogen-activated protein kinase/extracellular signal-regulated kinase, and extracellular signal-regulated kinase inhibition (
      • Abdel-Malek Z.
      • Swope V.B.
      • Pallas J.
      • Krug K.
      • Nordlund J.J.
      Mitogenic, melanogenic, and cAMP responses of cultured neonatal human melanocytes to commonly used mitogens.
      ,
      • Carreira S.
      • Goodall J.
      • Denat L.
      • Rodriguez M.
      • Nuciforo P.
      • Hoek K.S.
      • et al.
      Mitf regulation of Dia1 controls melanoma proliferation and invasiveness.
      ,
      • Cheli Y.
      • Giuliano S.
      • Botton T.
      • Rocchi S.
      • Hofman V.
      • Hofman P.
      • et al.
      Mitf is the key molecular switch between mouse or human melanoma initiating cells and their differentiated progeny.
      ,
      • Cheli Y.
      • Giuliano S.
      • Fenouille N.
      • Allegra M.
      • Hofman V.
      • Hofman P.
      • et al.
      Hypoxia and MITF control metastatic behaviour in mouse and human melanoma cells.
      ,
      • Hoek K.S.
      • Goding C.R.
      Cancer stem cells versus phenotype-switching in melanoma.
      ,
      • Kemper K.
      • de Goeje P.L.
      • Peeper D.S.
      • van Amerongen R.
      Phenotype switching: tumor cell plasticity as a resistance mechanism and target for therapy.
      ,
      • Konieczkowski D.J.
      • Johannessen C.M.
      • Abudayyeh O.
      • Kim J.W.
      • Cooper Z.A.
      • Piris A.
      • et al.
      A melanoma cell state distinction influences sensitivity to MAPK pathway inhibitors.
      ,
      • Roesch A.
      • Fukunaga-Kalabis M.
      • Schmidt E.C.
      • Zabierowski S.E.
      • Brafford P.A.
      • Vultur A.
      • et al.
      A temporarily distinct subpopulation of slow-cycling melanoma cells is required for continuous tumor growth.
      ,
      • Verfaillie A.
      • Imrichova H.
      • Atak Z.K.
      • Dewaele M.
      • Rambow F.
      • Hulselmans G.
      • et al.
      Decoding the regulatory landscape of melanoma reveals TEADS as regulators of the invasive cell state.
      ), and successful treatment requires the eradication of this subpopulation. The transcriptional control of these opposing states is being extensively studied. To date, however, little is known about the role of microRNAs (miRNAs or miRs) in this process. miRNAs are small non-protein-coding RNAs that negatively regulate gene expression posttranscriptionally. There are currently 2,588 potential human miRNAs recorded in the miRBase (released June 21, 2014), and they are predicted to regulate more than 60% of all genes (
      • Lewis B.P.
      • Burge C.B.
      • Bartel D.P.
      Conserved seed pairing, often flanked by adenosines, indicates that thousands of human genes are microRNA targets.
      ). miRNAs are now recognized as key components in cancer biology because they control tumor growth, invasion, angiogenesis, and immune evasion (
      • Hayes J.
      • Peruzzi P.P.
      • Lawler S.
      MicroRNAs in cancer: biomarkers, functions and therapy.
      ). Deregulated miRNA expression has been linked with a tumor progression in a variety of cancers including melanoma (
      • Croce C.M.
      Causes and consequences of microRNA dysregulation in cancer.
      ,
      • Mione M.
      • Bosserhoff A.
      MicroRNAs in melanocyte and melanoma biology.
      ). Transcription factor-miRNA relationships modulating proliferative and invasive programs in melanoma cell lines were recently characterized (
      • Bell R.E.
      • Khaled M.
      • Netanely D.
      • Schubert S.
      • Golan T.
      • Buxbaum A.
      • et al.
      Transcription factor/microRNA axis blocks melanoma invasion program by miR-211 targeting NUAK1.
      ). The MITF/miR-211/NUAK1 axis has been identified as an inhibitor of the invasive program, acting by preventing cell adhesion (
      • Bell R.E.
      • Khaled M.
      • Netanely D.
      • Schubert S.
      • Golan T.
      • Buxbaum A.
      • et al.
      Transcription factor/microRNA axis blocks melanoma invasion program by miR-211 targeting NUAK1.
      ).
      To further investigate the role of miRNAs in the maintenance of the invasive melanoma cell state, we combined miRNA and phenotypic profiling using established human melanoma cell lines. With wound scratch and Matrigel (Corning Matrigel Growth Factor Reduced Basement Membrane Matrix, Corning, Bagneux-sur-Loing, France) invasion assays, we identified two phenotypically different groups of melanoma cell lines. Three miRNAs overexpressed in proinvasive and promigratory melanoma cell lines, namely miR-125b, miR-100, and miR-199b-5p, were also highly expressed in melanomas of the invasive phenotype. Exploiting miRNA expression data and clinical data for The Cancer Genome Atlas (TCGA) melanoma cohort, we identified miR-125b as being significantly overexpressed in metastatic melanoma and linked to reduced patient survival. The inhibition of miR-125b decreased migratory and invasive potentials of aggressive melanoma cells in vitro. We show that neural precursor cell expressed developmentally down-regulated protein 9 (NEDD9) is a direct target of miR-125b, thereby modulating melanoma cell migration. Also, overexpression of transcription factor 4 (TCF4), an epithelial-mesenchymal transition-inducing transcription factor, induces MIR125B1 expression. Thus, miR-125b appears to regulate melanoma progression by promoting the establishment of the invasive cell state reminiscent of a pseudo–epithelial-mesenchymal transition (
      • Larue L.
      • Beermann F.
      Cutaneous melanoma in genetically modified animals.
      ).

      Results

      The miRNA profile of invasive and migratory melanoma cells

      We asked whether discriminative miRNA expression patterns can be assigned to the invasive and proliferative transcriptional cell states. To this end, we first selected phenotypically distinct human melanoma lines by performing Matrigel (Corning) invasion and wound scratch assays. Cell lines could be classified into two classes according to their invasive and migratory potential: WM793, WM1366, Lu1205, WM852, and A375M cells (defined as class 1) were more invasive and migratory than 501mel, T1, G1, and MNT-1 cells (class 2) (Figure 1a). It appears that the class 1 melanoma cell lines overexpressed miR-125b, -21, -100, and -199b and underexpressed miR-513a-5p, -185, and -211; the reverse was true for the class 2 melanoma cell lines (Figure 1b, see Supplementary Table S1 online). Three phenotypic groups, invasive, immune, and proliferative, were recently proposed for bulk melanomas (
      • Verfaillie A.
      • Imrichova H.
      • Atak Z.K.
      • Dewaele M.
      • Rambow F.
      • Hulselmans G.
      • et al.
      Decoding the regulatory landscape of melanoma reveals TEADS as regulators of the invasive cell state.
      ). To investigate a potential in vivo relevance for our in vitro-derived miRNA profile we quantified these miRNAs in the three groups of melanomas. The miRNAs that were underexpressed in functionally promigratory and proinvasive melanoma cell lines (miR-513a, -185, and -211) were overexpressed in proliferative bulk melanomas (Figure 1c); miRNAs that were overexpressed in promigratory and proinvasive cell lines were also significantly overexpressed in invasive bulk melanomas (miR-125b, -100, and -199b). However, miR-21 appeared to be overexpressed in immune bulk melanomas (Figure 1d). Thus, these miRNAs are characteristic of phenotypically different cell states in vitro and are also markers of invasive and proliferative cell states in vivo.
      Figure 1
      Figure 1miRNA expression profile of phenotypically different melanoma cells. (a) Migratory (wound scratch assay) and invasive (Matrigel assay, Corning) capacities of nine human melanoma cell lines. Class 1 melanoma cell lines (WM793, WM1366, Lu1205, WM852, and A375M) are more aggressive than class 2 melanoma cell lines (501mel, T1, G1, and MNT-1). (b) miRNA expression profiling of the nine melanoma cell lines showed an underexpression of miR-513a-5p, miR-185, and miR-211 and an overexpression of miR-125b, miR-21, miR-100, and miR-199-5p in aggressive melanoma cells (FC > 2, adj. P < 0.01). (c) Poorly aggressive miRNAs were assayed in in vivo melanoma samples (TCGA SKCM cohort). Melanoma samples were classified into invasive (Inv), immune (Imm), and proliferative (Pro) molecular phenotypes based on global mRNA expression according to (
      • Verfaillie A.
      • Imrichova H.
      • Atak Z.K.
      • Dewaele M.
      • Rambow F.
      • Hulselmans G.
      • et al.
      Decoding the regulatory landscape of melanoma reveals TEADS as regulators of the invasive cell state.
      ). miR-513a, miR-185, and miR-211 were overexpressed in the proliferative category of melanomas. Mann-Whitney test, error bars = 95% confidence interval, ∗P < 0.05, ∗∗P < 10–2, ∗∗∗P < 10–3, and ∗∗∗∗P < 10–4. (d) Aggressive miRNAs were quantified in melanoma samples (TCGA SKCM cohort). miR-125b-1, miR-21, miR-100, and miR-199b were overexpressed in the invasive category of melanomas. adj.P, adjusted P-value; AU, arbitrary unit; FC, fold change; Imm, immune; Inv, invasive; miR, microRNA; miRNA, microRNA; mRNA, messenger RNA; ns, not significant; Pro, proliferative; SKCM, skin cutaneous melanoma; TCGA, The Cancer Genome Atlas.

      miR-125b overexpression in human melanoma metastases is associated with poor survival

      To further prioritize miRNAs being involved in melanoma progression, we checked whether the miRNAs that are overexpressed in functionally proinvasive and promigratory melanoma cell lines are differently expressed in primary and metastatic melanoma samples. We observed that miR-125b was significantly overexpressed in metastatic compared with primary melanomas (Figure 2a) (fold change = 1.75, P = 3 × 10–4). Additionally, miR-125b overexpression was significantly associated with reduced overall survival in cases of metastatic melanoma (Figure 2b). miR-125b was the only miRNA that was overexpressed in functionally proinvasive and promigratory melanoma cell lines (Figure 1), overexpressed in metastatic melanoma patients, and significantly linked to reduced overall survival (see Supplementary Figure S1 online). The expression of miR-125b has been reported to be weaker in melanoma cell lines than in normal human epidermal melanocytes or normal human epidermal melanocytes (NHEM) (
      • Mione M.
      • Bosserhoff A.
      MicroRNAs in melanocyte and melanoma biology.
      ,
      • Kappelmann M.
      • Kuphal S.
      • Meister G.
      • Vardimon L.
      • Bosserhoff A.K.
      MicroRNA miR-125b controls melanoma progression by direct regulation of c-Jun protein expression.
      ), so we used quantitative real-time reverse transcriptase (QRT)-PCR to quantify miR-125b expression in 10 human melanoma cell lines and NHEMs from six different donors (Figure 2c): the various cell lines showed different levels of miR-125b expression. NHEM primary cell cultures all expressed moderate miR-125b levels, whereas melanoma cell lines expressed either lower or higher levels than NHEMs. There are two MIR125B genes, one on chromosome 11 (mir-125b-1) and the other one on chromosome 21 (mir-125b-2). The levels of mir-125b-1 and mir-125b-2 precursors were evaluated in TCGA melanoma samples and quantified in Lu1205 cells by QRT-PCR; mir-125b-1 was much more abundant than mir-125b-2 in both melanoma samples and in Lu1205 cells (see Supplementary Figure S2 online).
      Figure 2
      Figure 2Elevated miR-125b levels are associated with shorter overall survival and aggressiveness in vitro. (a) miR-125b expression levels (Mann Whitney test, P = 3 × 10–4) are 75% higher in metastatic than primary melanoma cells (TCGA SKCM). Error bars represent 95% confidence interval. Mann-Whitney test, ∗∗∗P < 10–3. (b) Survival analysis of metastatic melanoma patients based on high and low miR-125b levels (TCGA SKCM) using 91 patients in both cases. The Kaplan Meier curves for high and low miR-125b levels are significantly different (P < 0.05) as assessed by the Mantel-Cox test. (c) miR-125b is heterogeneously expressed in human melanoma cell lines and primary melanocyte cultures. dp, darkly pigmented; Met, metastatic; miR, microRNA; NHEM, normal human epidermal melanocyte; Prim, primary; RPMMM, reads per million miRNA mapped; SKCM, skin cutaneous melanoma; TCGA, The Cancer Genome Atlas.

      Inhibition of miR-125b reduces melanoma cell migration and invasion in vitro

      We tested whether miR-125b expression affects melanoma cell migration and invasion. We treated miR-125bhigh Lu1205 melanoma cells with miR-125b inhibitors (i.e., antagomiRs). After 72 hours, miR-125b expression levels were significantly reduced (Figure 3a), but the proliferative capacity of the cells remained unaltered as indicated by BrdU incorporation (Figure 3b). However, miR-125b inhibition significantly reduced the migratory and invasive capacities of Lu1205 melanoma cells according to wound scratch and Matrigel (Corning) invasion assays (Figure 3c and d). Similar results were obtained with two other miR125bhigh cell lines: WM852 and WM1366 (see Supplementary Figure S3 online and not shown). These findings indicate that the reduction of miR-125b levels in miR-125bhigh melanoma cells results in impaired migratory and invasive potentials without affecting cell proliferation.
      Figure 3
      Figure 3Inhibition of miR-125b reduces melanoma cell migration and invasion. (a) miR-125b levels in Lu1205 melanoma cells are significantly reduced after antagomiR treatment (quantitative real-time reverse transcription PCR). The control inhibitor was cel-miR-239b. (b) The proliferative capacity of Lu1205 melanoma cells, assessed using FACS-based BrdU incorporation assay, was unaltered after miR-125b inhibition. (c) Migratory capacity, assessed using wound healing assay (wound closure after 15 hours), of Lu1205 melanoma cells was significantly reduced upon miR-125b inhibition. (d) The Matrigel (Corning) invasion capacity of Lu1205 melanoma cells was significantly reduced upon miR-125b inhibition (Mann Whitney test, P = 0.033). ∗P < 0.05, ∗∗P < 10−2. Ctrl, control; ns, not significant; r.u., relative unit.

      miR-125b targets NEDD9 and affects cell migration

      We applied the following target prediction pipeline (see Supplementary Figure S4a online) to identify miR-125b targets involved in cell migration and/or invasion: messenger RNA (mRNA) targets should be (i) inversely correlated to miR-125b expression (rPearson< –0.75) and significantly differentially expressed between cell line classes 1 and 2 (fold change > 2, adjusted P < 0.05), (ii) predicted by the miRNA Data Integration Portal (i.e., mirDIP, a sequence-based prediction approach; Jurisica Lab, Ontario Cancer Institute, Princess Margaret Hospital/UHN), and (iii) suggested by Ingenuity Pathway Analysis (Qiagen, Hilden, Germany) to be involved in cell migration and/or invasion. Using this approach, we identified 13 potential targets (see Supplementary Table S2 online). We concentrated on NEDD9 (Figure S5a), recently implicated in melanoma cell migration and invasion (
      • Domingues M.J.
      • Rambow F.
      • Job B.
      • Papon L.
      • Liu W.
      • Larue L.
      • et al.
      Beta-catenin inhibitor ICAT modulates the invasive motility of melanoma cells.
      ). We used the pmirGLO dual-luciferase assay to validate the predicted miR-125b target site in the NEDD9 3′ untranslated region as follows: the wild-type and a mutated form of the miR-125b binding site were inserted into the reporter system. Upon treatment with miR-125b mimics, we detected a significant decrease in luciferase activity for the wild-type condition but not for the mutant as expected (Figure 4a). Next, we tested whether miR-125b overexpression affected the production of NEDD9 mRNA and protein. We transfected 501mel melanoma cells (miR-125blow) with miR-125b mimics for 48 hours; NEDD9 mRNA and protein levels decreased (Figure 4b). Conversely, we inhibited miR-125b expression in Lu1205 cells (miR-125bhigh), and both NEDD9 mRNA and protein levels increased (Figure 4c). Thus, NEDD9 is a target of miR-125b. We overexpressed miR-125b in 501mel cells and monitored single-cell migration using videomicroscopy. Upon miR-125b overexpression, NEDD9 mRNA levels dropped, and the migration speed of 501mel cells increased significantly. The same effect was observed when Lu1205 cells were treated with a small interfering RNA directed against NEDD9 (Figure 4d).
      Figure 4
      Figure 4NEDD9 is a miR-125b target involved in melanoma cell migration. (a) Luciferase activity assay validates predicted miR-125b binding site in the NEDD9 3′ untranslated region. The predicted binding site (WT) and a mutated version (MUT) were inserted into the pmiRGLO system. The mutated nucleotides are in red. Upon miR-125b overexpression, luciferase activity was significantly reduced by the presence of the WT sequence. (b) NEDD9 mRNA and protein levels were reduced upon miR-125b overexpression in 501mel melanoma cells, expressing miR-125b at a low level (miR-125blow). (c) NEDD9 mRNA and protein levels were increased upon miR-125b inhibition in Lu1205 melanoma cells, expressing miR-125b at a high level (miR-125bhigh). (d) Single-cell migration speed of miR-125blow 501mel cells was increased and NEDD9 mRNA levels decreased upon miR-125b overexpression. Single-cell migration speed of miR-125bhigh Lu1205 cells was increased and NEDD9 mRNA levels decreased upon siRNA mediated knockdown of NEDD9. Mann Whitney test, error bars = 95% confidence interval; P < 0.05, ∗∗P < 10–2, and ∗∗∗P < 10–3. h, hours; luc, luciferase; miR, microRNA; mRNA, messenger RNA; MUT, mutated form; ns, not significant; NEDD9, neural precursor cell expressed developmentally down-regulated protein 9; RQ, relative quantity; ru, relative unit; siRNA, small interfering RNA; UTR, untranslated region; WT, wild type.

      TCF4 is an upstream regulator of MIR125B1

      To identify transcription factors regulating miR-125b, we looked for those (i) for which the abundance of mRNA (identified by Ingenuity Pathway Analysis) coincided with miR-125b levels in our melanoma cell lines and (ii) were differentially expressed between the four promigratory/proinvasive and five proliferative cell lines described in Figure 1. Forty transcription factors fitted these criteria (see Supplementary Figure S4b and Supplementary Table S3 online). Their mRNA levels were then compared with miR-125b levels in the TCGA melanoma cohort (see Supplementary Table S4 online). The basic helix-loop-helix TCF4 (called ITF2 or E2-2) and ZEB1, two transcription factor proteins involved in epithelial-mesenchymal transition transcription factors were identified among the top five candidates (based on P-value and r > 0.4). TCF4 induces miR-125b expression in breast cancer cell lines (
      • Liu Z.
      • Liu H.
      • Desai S.
      • Schmitt D.C.
      • Zhou M.
      • Khong H.T.
      • et al.
      miR-125b functions as a key mediator for snail-induced stem cell propagation and chemoresistance.
      ) and may act upstream from ZEB1 (
      • Sobrado V.R.
      • Moreno-Bueno G.
      • Cubillo E.
      • Holt L.J.
      • Nieto M.A.
      • Portillo F.
      • et al.
      The class I bHLH factors E2-2A and E2-2B regulate EMT.
      ). We therefore studied TCF4 further as a potential upstream regulator of miR-125b in melanoma.
      TCF4 and miR-125b abundance correlated significantly (r = 0.82, P = 3.5 × 10–3) in our melanoma cell lines (see Supplementary Figure S5b online). In addition, we investigated histone H3 acetyl Lys27 (H3K27ac) chromatin immunoprecipitation sequencing (i.e., ChIP-seq) data for differences in active chromatin states between invasive and proliferative melanoma short-term cultures (
      • Verfaillie A.
      • Imrichova H.
      • Atak Z.K.
      • Dewaele M.
      • Rambow F.
      • Hulselmans G.
      • et al.
      Decoding the regulatory landscape of melanoma reveals TEADS as regulators of the invasive cell state.
      ). The putative promoter region of MIR125B1 showed more H3K27ac peaks in the invasive than the proliferative cell state (see Supplementary Figure S6a online). MM057 melanoma cells, initially termed proliferative, show some H3K27ac marks on miR-125b and NEDD9 genes. These melanoma cells belong to a potential intermediate cell state, because they express both proliferative and invasive genes such as TCF4 and ZEB1, for example (
      • Verfaillie A.
      • Imrichova H.
      • Atak Z.K.
      • Dewaele M.
      • Rambow F.
      • Hulselmans G.
      • et al.
      Decoding the regulatory landscape of melanoma reveals TEADS as regulators of the invasive cell state.
      ). Up to 15% of profiled melanoma cell lines might fall into this intermediate state (
      • Widmer D.S.
      • Cheng P.F.
      • Eichhoff O.M.
      • Belloni B.C.
      • Zipser M.C.
      • Schlegel N.C.
      • et al.
      Systematic classification of melanoma cells by phenotype-specific gene expression mapping.
      ).
      To test for the involvement of TCF4 in the regulation of the MIR125B1 promoter, we stably infected proliferative MM034 melanoma cells with a TCF4 lentivirus. The transduction resulted in significant overexpression of TCF4 as shown by QRT-PCR and Western blot analysis (Figure 5a and b). Also, MM034 cells overexpressing TCF4 showed an up-regulation of mir-125b-1 expression and a reduction in NEDD9 mRNA and protein levels (Figure 5a and b). The inverse expression of TCF4 and NEDD9 in invasive and proliferative melanoma cells was also confirmed in the independent melanoma short-term culture data set (see Supplementary Figure S6b and c). Altogether, the results suggest the existence of a signaling cascade in which TCF4 regulates miR-125b, which in turn targets NEDD9, leading to increased migration and invasion of melanoma cells (Figure 5c).
      Figure 5
      Figure 5TCF4 induces mir-125b-1 expression. MM034 melanoma cells were stably infected with a lentivirus for TCF4 (MM034_TCF). (a) Semi-quantitative reverse transcription-PCR analysis shows up-regulation of TCF4 and mir-125b-1 and down-regulation of NEDD9 in cells stably overexpressing TCF4 relative to MM034 wild-type cells (MM034_WT). (b) Western blot analysis showing the overexpression of TCF4 after viral infection and the reduction of the amount of NEDD9 protein. Actin was used as loading control. (c) Schematic model for the TCF4/miR-125b/NEDD9 cascade. TCF4 drives miR-125b-1 expression, which upon maturation targets the 3′ UTR of NEDD9, leading to a reduction of NEDD9 transcript and protein levels and an increase in melanoma cell migration and invasion. FC, fold change; miR, microRNA; NEDD9, neural precursor cell expressed developmentally down-regulated protein 9; TCF4, transcription factor 4; UTR, untranslated region; WT, wild type.

      Discussion

      Reciprocal gene expression programs defined two melanoma subpopulations that were proliferative and invasive in in vitro assays (
      • Hoek K.S.
      • Eichhoff O.M.
      • Schlegel N.C.
      • Dobbeling U.
      • Kobert N.
      • Schaerer L.
      • et al.
      In vivo switching of human melanoma cells between proliferative and invasive states.
      ,
      • Hoek K.S.
      • Schlegel N.C.
      • Brafford P.
      • Sucker A.
      • Ugurel S.
      • Kumar R.
      • et al.
      Metastatic potential of melanomas defined by specific gene expression profiles with no BRAF signature.
      ). Proliferative melanoma cells have strong melanocyte-specific MITF (MITF-M) expression signatures and generally respond to combination therapy, whereas invasive melanoma cells have weak MITF-M expression, greater resistance to BRAF and mitogen-activated protein kinase/extracellular signal-regulated kinase inhibition, and apparent stem-like properties (
      • Hoek K.S.
      • Eichhoff O.M.
      • Schlegel N.C.
      • Dobbeling U.
      • Kobert N.
      • Schaerer L.
      • et al.
      In vivo switching of human melanoma cells between proliferative and invasive states.
      ,
      • Muller J.
      • Krijgsman O.
      • Tsoi J.
      • Robert L.
      • Hugo W.
      • Song C.
      • et al.
      Low MITF/AXL ratio predicts early resistance to multiple targeted drugs in melanoma.
      ). Melanoma tissue samples also display such opposing gene expression signatures (
      • Verfaillie A.
      • Imrichova H.
      • Atak Z.K.
      • Dewaele M.
      • Rambow F.
      • Hulselmans G.
      • et al.
      Decoding the regulatory landscape of melanoma reveals TEADS as regulators of the invasive cell state.
      ). We investigated the miRNA profile of invasive melanoma cells, because promising ongoing clinical trials have shown that miRNAs play a key role in cancer biology (
      • Hayes J.
      • Peruzzi P.P.
      • Lawler S.
      MicroRNAs in cancer: biomarkers, functions and therapy.
      ). We performed miRNA and phenotypic profiling in parallel in nine melanoma cell lines: four proliferative and five invasive. Microarray analysis showed miR-125b, -21, -100, and -199b-5p to be overexpressed in invasive melanoma cells and miR-513a-5p, -185, and -211 to be underexpressed. We investigated the miRNome of invasive cells, focusing on miR-125b, -21, -100, and -199b-5p. Data for 375 melanoma patients (TCGA skin cutaneous melanoma cohort) showed that miR-125b, -100, and miR-199b-5p were significantly overexpressed in invasive melanomas (
      • Verfaillie A.
      • Imrichova H.
      • Atak Z.K.
      • Dewaele M.
      • Rambow F.
      • Hulselmans G.
      • et al.
      Decoding the regulatory landscape of melanoma reveals TEADS as regulators of the invasive cell state.
      ), consistent with the recent transcriptomic classification of melanomas by the TCGA Consortium showing significant up-regulation of miR-125b, -100, and -199b-5p in the MITF-low subclass (
      Cancer Genome Atlas Network
      Genomic classification of cutaneous melanoma.
      ). The expression of genes associated with pigmentation and epithelial expression characteristics, including several MITF target genes (a hallmark of invasive melanoma cells), is weak in MITF-low tumors. The overexpression of miR-125b in melanocytes leads to the down-regulation of pigmentation genes, including tyrosinase (TYR) and dopachrome tautomerase (DCT), independently of MITF (
      • Kim K.H.
      • Bin B.H.
      • Kim J.
      • Dong S.E.
      • Park P.J.
      • Choi H.
      • et al.
      Novel inhibitory function of miR-125b in melanogenesis.
      ). miR-125b was the only miRNA in our dataset significantly overexpressed in metastatic melanoma and associated with poor patient survival. Human miR-125b is ubiquitous in normal tissue and dysregulated in diverse tumors (
      • Huang K.
      • Dong S.
      • Li W.
      • Xie Z.
      The expression and regulation of microRNA-125b in cancers.
      ). miR-125b may be oncogenic or tumor suppressive, depending on tumor type (
      • Banzhaf-Strathmann J.
      • Edbauer D.
      Good guy or bad guy: the opposing roles of microRNA 125b in cancer.
      ,
      • Sun Y.M.
      • Lin K.Y.
      • Chen Y.Q.
      Diverse functions of miR-125 family in different cell contexts.
      ). miR-125b is a key regulator of hair follicle stem cells, conferring “oncomiR addiction” in an early-stage squamous cell carcinoma model by favoring a cancer stem cell-like transcription program (
      • Zhang L.
      • Ge Y.
      • Fuchs E.
      miR-125b can enhance skin tumor initiation and promote malignant progression by repressing differentiation and prolonging cell survival.
      ,
      • Zhang L.
      • Stokes N.
      • Polak L.
      • Fuchs E.
      Specific microRNAs are preferentially expressed by skin stem cells to balance self-renewal and early lineage commitment.
      ).
      miR-125b is overexpressed in highly aggressive C8161 melanoma cells (
      • Greenberg E.
      • Hershkovitz L.
      • Itzhaki O.
      • Hajdu S.
      • Nemlich Y.
      • Ortenberg R.
      • et al.
      Regulation of cancer aggressive features in melanoma cells by microRNAs.
      ) and is one of the top 50 miRNA predictors of metastatic melanoma progression (
      • Li Y.
      • Krahn J.M.
      • Flake G.P.
      • Umbach D.M.
      • Li L.
      Toward predicting metastatic progression of melanoma based on gene expression data.
      ). It has been identified as part of an overexpressed miRNA profile in TGFB1+ invasive melanoma cell lines (
      • Hwang H.W.
      • Baxter L.L.
      • Loftus S.K.
      • Cronin J.C.
      • Trivedi N.S.
      • Borate B.
      • et al.
      Distinct microRNA expression signatures are associated with melanoma subtypes and are regulated by HIF1A.
      ), and its levels were inversely correlated with the proliferative mRNA expression signature in 88 melanoma cell lines (
      • Bell R.E.
      • Khaled M.
      • Netanely D.
      • Schubert S.
      • Golan T.
      • Buxbaum A.
      • et al.
      Transcription factor/microRNA axis blocks melanoma invasion program by miR-211 targeting NUAK1.
      ). We found that transforming growth factor-β increased miR-125b levels by up to 50% in 501mel cells (see Supplementary Figure S7a online). Hypoxia has also been shown to induce an invasive melanoma phenotype (
      • Schlegel N.C.
      • von Planta A.
      • Widmer D.S.
      • Dummer R.
      • Christofori G.
      PI3K signalling is required for a TGFbeta-induced epithelial-mesenchymal-like transition (EMT-like) in human melanoma cells.
      ). Consistent with this observation, miR-125b is induced by hypoxia in endothelial and carcinoma cells (
      • Muramatsu F.
      • Kidoya H.
      • Naito H.
      • Sakimoto S.
      • Takakura N.
      microRNA-125b inhibits tube formation of blood vessels through translational suppression of VE-cadherin.
      ,
      • Kulshreshtha R.
      • Ferracin M.
      • Wojcik S.E.
      • Garzon R.
      • Alder H.
      • Agosto-Perez F.J.
      • et al.
      A microRNA signature of hypoxia.
      ). The invasive melanoma state was recently linked to treatment resistance.
      • Vergani E.
      • Di Guardo L.
      • Dugo M.
      • Rigoletto S.
      • Tragni G.
      • Ruggeri R.
      • et al.
      Overcoming melanoma resistance to vemurafenib by targeting CCL2-induced miR-34a, miR-100 and miR-125b.
      showed that targeting miR-125b through chemokine (C-C motif) ligand 2 (CCL2) would overcome resistance to BRAF inhibition. We analyzed the correlation between miR-125b expression and resistance to BRAF inhibition in 28 melanoma cell lines from the Cancer Cell Line Encyclopedia cohort. We found that BRAF inhibitor resistance was significantly correlated with high levels of miR-125b (see Supplementary Figure S7b).
      Consistent with these findings, we showed that miR-125b inhibition decreases the potential for migration and invasion in vitro. Gain- and loss-of-function assays showed partial down-regulation of NEDD9 by miR-125b. NEDD9 belongs to the cellular apoptosis susceptibility (CAS) family of scaffold proteins involved in the mesenchymal movement of melanoma cells (
      • Sanz-Moreno V.
      • Gadea G.
      • Ahn J.
      • Paterson H.
      • Marra P.
      • Pinner S.
      • et al.
      Rac activation and inactivation control plasticity of tumor cell movement.
      ). Some metastatic melanoma samples have been reported to display strong NEDD9 expression (
      • Kim M.
      • Gans J.D.
      • Nogueira C.
      • Wang A.
      • Paik J.H.
      • Feng B.
      • et al.
      Comparative oncogenomics identifies NEDD9 as a melanoma metastasis gene.
      ), but studies in a murine mammary tumor model showed NEDD9 deficiency to be associated with tumor hyperaggressiveness (
      • Singh M.K.
      • Izumchenko E.
      • Klein-Szanto A.J.
      • Egleston B.L.
      • Wolfson M.
      • Golemis E.A.
      Enhanced genetic instability and dasatinib sensitivity in mammary tumor cells lacking NEDD9.
      ). NEDD9 down-regulation by inhibitor of β-catenin and TCF increases the motility, invasiveness, and metastatic potential of Lu1205 melanoma cells (
      • Domingues M.J.
      • Rambow F.
      • Job B.
      • Papon L.
      • Liu W.
      • Larue L.
      • et al.
      Beta-catenin inhibitor ICAT modulates the invasive motility of melanoma cells.
      ), suggesting a complex role for NEDD9 in tumor formation and metastasis.
      • Nyholm A.M.
      • Lerche C.M.
      • Manfe V.
      • Biskup E.
      • Johansen P.
      • Morling N.
      • et al.
      miR-125b induces cellular senescence in malignant melanoma.
      suggested that miR-125b acted as a tumor suppressor, consistent with contrasting roles for miR-125b in early tumor development and melanoma progression. miR-125b overexpression in Mel-Juso human melanoma cells leads to cell cycle arrest in the G0/G1 phase transition and an up-regulation of senescence markers (
      • Glud M.
      • Manfe V.
      • Biskup E.
      • Holst L.
      • Dirksen A.M.
      • Hastrup N.
      • et al.
      MicroRNA miR-125b induces senescence in human melanoma cells.
      ,
      • Nyholm A.M.
      • Lerche C.M.
      • Manfe V.
      • Biskup E.
      • Johansen P.
      • Morling N.
      • et al.
      miR-125b induces cellular senescence in malignant melanoma.
      ). The inhibition of melanoma cell proliferation by miR-125b after the exogenous overexpression of miR-125b in miR-125b–low Melim and Mel-Juso melanoma cell lines has been reported (
      • Kappelmann M.
      • Kuphal S.
      • Meister G.
      • Vardimon L.
      • Bosserhoff A.K.
      MicroRNA miR-125b controls melanoma progression by direct regulation of c-Jun protein expression.
      ). Jun protein levels were also low, suggesting possible regulation of c-Jun–targeted genes involved in the cell cycle, proliferation, and differentiation by miR-125b (
      • Kappelmann M.
      • Kuphal S.
      • Meister G.
      • Vardimon L.
      • Bosserhoff A.K.
      MicroRNA miR-125b controls melanoma progression by direct regulation of c-Jun protein expression.
      ). miR-125b levels were generally lower in melanoma cell lines than in NHEMs, but we found a previously unappreciated group of melanoma cell lines with higher miR-125b levels than six different NHEM donors. The effect of miR-125b on proliferation seems to be determined by the existing endogenous miR-125b levels in the cell. Cell proliferation may be down-regulated only if a threshold level of miR-125b is exceeded.
      The high miR-125b levels in TCGA melanoma metastases also suggest a role in melanoma progression. We provide direct evidence for a complex role of miR-125b in melanoma progression. Gene expression and H3K27ac chromatin immunoprecipitation sequencing data suggested that TCF4 and miR-125b were characteristic of invasive cells. Indeed, stable TCF4 overexpression in MM034 melanoma cells increased mir-125b-1 levels and reduced NEDD9 mRNA and protein levels. TCF4, encoded by a gene on chromosome 18q21.1, recognizes E-box motifs (CANNTG). miR-125b has been identified as a key mediator in Snail-induced breast cancer-stem-cell-enrichment and chemoresistance downstream from TCF4 (
      • Liu Z.
      • Liu H.
      • Desai S.
      • Schmitt D.C.
      • Zhou M.
      • Khong H.T.
      • et al.
      miR-125b functions as a key mediator for snail-induced stem cell propagation and chemoresistance.
      ). TCF4 overexpression induces full epithelial-mesenchymal transition in epithelial Madin-Darby canine kidney epithelial (MDCK) cells (
      • Sobrado V.R.
      • Moreno-Bueno G.
      • Cubillo E.
      • Holt L.J.
      • Nieto M.A.
      • Portillo F.
      • et al.
      The class I bHLH factors E2-2A and E2-2B regulate EMT.
      ) and contributes to phenotype switching in melanoma, as part of the invasive melanoma signature (n = 45) (
      • Eichhoff O.M.
      • Weeraratna A.
      • Zipser M.C.
      • Denat L.
      • Widmer D.S.
      • Xu M.
      • et al.
      Differential LEF1 and TCF4 expression is involved in melanoma cell phenotype switching.
      ,
      • Widmer D.S.
      • Cheng P.F.
      • Eichhoff O.M.
      • Belloni B.C.
      • Zipser M.C.
      • Schlegel N.C.
      • et al.
      Systematic classification of melanoma cells by phenotype-specific gene expression mapping.
      ). This study shows a role for miR-125b in melanoma progression, in which it is regulated by TCF4 and targets NEDD9. These findings shed light on the way in which melanoma cells adopt an invasive cell state by adapting their miRNome. This phenomenon seems to be extremely difficult to overcome to ensure effective antimelanoma therapy.

      Materials and Methods

      RNA extraction

      Total RNA was extracted from melanoma cell lines using the TRIzol reagent by standard method (Invitrogen, Carlsbad, CA). Total RNA was purified using the RNA cleanup and concentration kit (RNeasy Micro Kit, Qiagen, Hilden, Germany).

      miRNA sample preparation and hybridization

      Each sample was prepared according to Agilent’s (CA) miRNA Microarray System protocol (see Supplementary Materials online).

      mRNA sample preparation and hybridization

      Each sample was prepared according to Agilent’s mRNA Microarray System protocol (see Supplementary Materials).

      Bioinformatics and statistics analyses miRNA and mRNA assay

      We used human microRNA microarrays version 3.0 (Agilent), which contain 851 human miRNAs and mRNA microarrays (Agilent) that target 19,596 RNA genes (Gene Expression Omnibus number: GSE67638). The Cy3 median signal was used for miRNA and mRNA analyses, using functions and packages collected in the Bioconductor project (open source software for bioinformatics developed at the Hutchinson Cancer Research Center, USA). Clusters were computed using the “dist” function of R software, using the “euclidian” method as the distance measure. The hierarchical clustering was performed with the “hclust” function of R, using the distance matrix previously computed and the ward method (see Supplementary Materials).

      Statistical analysis

      To study differentially expressed probes, we first estimated the fold differences and standard errors between two groups of samples by fitting a linear model for each probe with the “lmFit” function of the Linear Models for Microarray Data package (see
      • Ritchie M.E.
      • Phipson B.
      • Wu D.
      • Hu Y.
      • Law C.W.
      • Shi W.
      • et al.
      limma powers differential expression analyses for RNA-sequencing and microarray studies.
      and Supplementary Materials).

      QRT-PCR to study miRNA

      Total RNA was isolated using the miRNeasy Mini Kit (Qiagen). Complementary DNA was synthesized using the miScript II Reverse Transcription Kit (Qiagen) and real-time PCR, with the miScript SYBR Green PCR kit (Qiagen) (see Supplementary Materials).

      QRT-PCR to study mRNA

      Total RNA was isolated using the miRNeasy Mini Kit (Qiagen). Complementary DNA was synthesized using the miScript II Reverse Transcription Kit (Qiagen) and real-time PCR with the iTaq universal Sybrgreen supermix (Bio-Rad, CA) (see Supplementary Materials).

      Wound scratch assay

      Wound scratch assay was performed as described previously (
      • Gallagher S.J.
      • Rambow F.
      • Kumasaka M.
      • Champeval D.
      • Bellacosa A.
      • Delmas V.
      • et al.
      Beta-catenin inhibits melanocyte migration but induces melanoma metastasis.
      ) (see Supplementary Materials).

      pmiRGLO assay

      Lu1205 cells were transfected, using Lipofectamine 2000 (Invitrogen), with 50 nmol/L of miR-125b mimics (Dharmacon, CO) and 40 ng of various pmiRGLO (Promega, WI) dual-luciferase miRNA target vectors (see Supplementary Materials).

      Time-lapse phase-contrast microscopy and tracking assays

      Single-cell tracking was performed as described previously (
      • Domingues M.J.
      • Rambow F.
      • Job B.
      • Papon L.
      • Liu W.
      • Larue L.
      • et al.
      Beta-catenin inhibitor ICAT modulates the invasive motility of melanoma cells.
      ) (see Supplementary Materials).

      BrdU incorporation assay

      Lu1205 cells were seeded on six-well plates (5 × 105). Twenty-four hours after seeding, cells were incubated with 10 mmol/L of BrdU (BD Biosciences, CA) for 1 hour (see Supplementary Materials).

      Western blotting

      Mouse monoclonal anti-TCF4 (Sigma-Aldrich, Lyon, France, SAB1412658, 1/3000), mouse monoclonal anti-NEDD9 (clone 2G9, 1/1000), and mouse monoclonal anti–β-actin (AC-15) A5441 (Sigma, 1/7500) were used (see Supplementary Materials).

      Cell culture

      Cell lines were selected and grown according to their molecular and phenotypic characteristics as well as their relevance as previously shown (
      • Rambow F.
      • Job B.
      • Petit V.
      • Gesbert F.
      • Delmas V.
      • Seberg H.
      • et al.
      New functional signatures for understanding melanoma biology from tumor cell lineage-specific analysis.
      ) (see Supplementary Materials).

      Mimic and miR-inhibitor treatment

      Melanoma cells were transfected with 50 nmol/L of miR-125b mimics or inhibitors (Dharmacon) (see Supplementary Material).

      Invasion assay

      Matrigel (Corning) invasion assays were performed as described previously (
      • Rambow F.
      • Job B.
      • Petit V.
      • Gesbert F.
      • Delmas V.
      • Seberg H.
      • et al.
      New functional signatures for understanding melanoma biology from tumor cell lineage-specific analysis.
      ) (see Supplementary Materials).

      miR-125b target prediction

      Potential targets of miR-125b were predicted using the miRNA Data Integration Portal algorithm (Jurisica Lab, Ontario Cancer Institute, Princess Margaret Hospital/UHN; see Supplementary Table S2 and Supplementary Materials).

      miR-125b regulator prediction

      We used Pearson correlation to identify transcription factors that were coexpressed with miR-125b (see Supplementary Table S3 and Supplementary Materials).

      Lentiviral transduction

      The TCF4 lentiviral construct was obtained by inserting the coding sequence of TCF4 in a lentiviral vector. Lentiviruses were produced by transient transfection of 293T cells, and supernatants were concentrated. MM034 cells were transduced and then harvested before molecular analysis (see Supplementary Materials).

      Data mining

      mRNA and miRNA expression data (level 3) were downloaded from the TCGA portal for 375 melanoma patients (see Supplementary Materials).

      Conflict of Interest

      The authors state no conflict of interest.

      Acknowledgments

      We thank the team of the imaging facility at Institut Curie for their support. This work was supported by the Ligue Nationale Contre le Cancer et comité de l’Oise, INCa, Pair Melanoma, and Cancéropole IdF, and is under the program “Investissements d’Avenir” launched by the French Government and implemented by ANR Labex CelTisPhyBio (ANR-11-LBX-0038 and ANR-10-IDEX-0001-02 PSL). We thank A. Rogiers for his assistance with the western blot analysis.

      Supplementary Material

      References

        • Abdel-Malek Z.
        • Swope V.B.
        • Pallas J.
        • Krug K.
        • Nordlund J.J.
        Mitogenic, melanogenic, and cAMP responses of cultured neonatal human melanocytes to commonly used mitogens.
        J Cell Physiol. 1992; 150: 416-425
        • Banzhaf-Strathmann J.
        • Edbauer D.
        Good guy or bad guy: the opposing roles of microRNA 125b in cancer.
        Cell Commun Signal. 2014; 12: 30
        • Bell R.E.
        • Khaled M.
        • Netanely D.
        • Schubert S.
        • Golan T.
        • Buxbaum A.
        • et al.
        Transcription factor/microRNA axis blocks melanoma invasion program by miR-211 targeting NUAK1.
        J Invest Dermatol. 2014; 134: 441-451
        • Cancer Genome Atlas Network
        Genomic classification of cutaneous melanoma.
        Cell. 2015; 161: 1681-1696
        • Carreira S.
        • Goodall J.
        • Denat L.
        • Rodriguez M.
        • Nuciforo P.
        • Hoek K.S.
        • et al.
        Mitf regulation of Dia1 controls melanoma proliferation and invasiveness.
        Genes Dev. 2006; 20: 3426-3439
        • Cheli Y.
        • Giuliano S.
        • Botton T.
        • Rocchi S.
        • Hofman V.
        • Hofman P.
        • et al.
        Mitf is the key molecular switch between mouse or human melanoma initiating cells and their differentiated progeny.
        Oncogene. 2011; 30: 2307-2318
        • Cheli Y.
        • Giuliano S.
        • Fenouille N.
        • Allegra M.
        • Hofman V.
        • Hofman P.
        • et al.
        Hypoxia and MITF control metastatic behaviour in mouse and human melanoma cells.
        Oncogene. 2012; 31: 2461-2470
        • Croce C.M.
        Causes and consequences of microRNA dysregulation in cancer.
        Nat Rev Genet. 2009; 10: 704-714
        • Domingues M.J.
        • Rambow F.
        • Job B.
        • Papon L.
        • Liu W.
        • Larue L.
        • et al.
        Beta-catenin inhibitor ICAT modulates the invasive motility of melanoma cells.
        Cancer Res. 2014; 74: 1983-1995
        • Eichhoff O.M.
        • Weeraratna A.
        • Zipser M.C.
        • Denat L.
        • Widmer D.S.
        • Xu M.
        • et al.
        Differential LEF1 and TCF4 expression is involved in melanoma cell phenotype switching.
        Pigment Cell Melanoma Res. 2011; 24: 631-642
        • Gallagher S.J.
        • Rambow F.
        • Kumasaka M.
        • Champeval D.
        • Bellacosa A.
        • Delmas V.
        • et al.
        Beta-catenin inhibits melanocyte migration but induces melanoma metastasis.
        Oncogene. 2013; 32: 2230-2238
        • Glud M.
        • Manfe V.
        • Biskup E.
        • Holst L.
        • Dirksen A.M.
        • Hastrup N.
        • et al.
        MicroRNA miR-125b induces senescence in human melanoma cells.
        Melanoma Res. 2011; 21: 253-256
        • Greenberg E.
        • Hershkovitz L.
        • Itzhaki O.
        • Hajdu S.
        • Nemlich Y.
        • Ortenberg R.
        • et al.
        Regulation of cancer aggressive features in melanoma cells by microRNAs.
        PLoS One. 2011; 6: e18936
        • Hayes J.
        • Peruzzi P.P.
        • Lawler S.
        MicroRNAs in cancer: biomarkers, functions and therapy.
        Trends Mol Med. 2014; 20: 460-469
        • Hoek K.S.
        • Eichhoff O.M.
        • Schlegel N.C.
        • Dobbeling U.
        • Kobert N.
        • Schaerer L.
        • et al.
        In vivo switching of human melanoma cells between proliferative and invasive states.
        Cancer Res. 2008; 68: 650-656
        • Hoek K.S.
        • Goding C.R.
        Cancer stem cells versus phenotype-switching in melanoma.
        Pigment Cell Melanoma Res. 2010; 23: 746-759
        • Hoek K.S.
        • Schlegel N.C.
        • Brafford P.
        • Sucker A.
        • Ugurel S.
        • Kumar R.
        • et al.
        Metastatic potential of melanomas defined by specific gene expression profiles with no BRAF signature.
        Pigment Cell Res. 2006; 19: 290-302
        • Huang K.
        • Dong S.
        • Li W.
        • Xie Z.
        The expression and regulation of microRNA-125b in cancers.
        Acta Biochim Biophys Sin (Shanghai). 2013; 45: 803-805
        • Hwang H.W.
        • Baxter L.L.
        • Loftus S.K.
        • Cronin J.C.
        • Trivedi N.S.
        • Borate B.
        • et al.
        Distinct microRNA expression signatures are associated with melanoma subtypes and are regulated by HIF1A.
        Pigment Cell Melanoma Res. 2014; 27: 777-787
        • Kappelmann M.
        • Kuphal S.
        • Meister G.
        • Vardimon L.
        • Bosserhoff A.K.
        MicroRNA miR-125b controls melanoma progression by direct regulation of c-Jun protein expression.
        Oncogene. 2013; 32: 2984-2991
        • Kemper K.
        • de Goeje P.L.
        • Peeper D.S.
        • van Amerongen R.
        Phenotype switching: tumor cell plasticity as a resistance mechanism and target for therapy.
        Cancer Res. 2014; 74: 5937-5941
        • Kim K.H.
        • Bin B.H.
        • Kim J.
        • Dong S.E.
        • Park P.J.
        • Choi H.
        • et al.
        Novel inhibitory function of miR-125b in melanogenesis.
        Pigment Cell Melanoma Res. 2014; 27: 140-144
        • Kim M.
        • Gans J.D.
        • Nogueira C.
        • Wang A.
        • Paik J.H.
        • Feng B.
        • et al.
        Comparative oncogenomics identifies NEDD9 as a melanoma metastasis gene.
        Cell. 2006; 125: 1269-1281
        • Konieczkowski D.J.
        • Johannessen C.M.
        • Abudayyeh O.
        • Kim J.W.
        • Cooper Z.A.
        • Piris A.
        • et al.
        A melanoma cell state distinction influences sensitivity to MAPK pathway inhibitors.
        Cancer Discov. 2014; 4: 816-827
        • Kulshreshtha R.
        • Ferracin M.
        • Wojcik S.E.
        • Garzon R.
        • Alder H.
        • Agosto-Perez F.J.
        • et al.
        A microRNA signature of hypoxia.
        Mol Cell Biol. 2007; 27: 1859-1867
        • Larue L.
        • Beermann F.
        Cutaneous melanoma in genetically modified animals.
        Pigment Cell Res. 2007; 20: 485-497
        • Lewis B.P.
        • Burge C.B.
        • Bartel D.P.
        Conserved seed pairing, often flanked by adenosines, indicates that thousands of human genes are microRNA targets.
        Cell. 2005; 120: 15-20
        • Li Y.
        • Krahn J.M.
        • Flake G.P.
        • Umbach D.M.
        • Li L.
        Toward predicting metastatic progression of melanoma based on gene expression data.
        Pigment Cell Melanoma Res. 2015; 28: 453-463
        • Liu Z.
        • Liu H.
        • Desai S.
        • Schmitt D.C.
        • Zhou M.
        • Khong H.T.
        • et al.
        miR-125b functions as a key mediator for snail-induced stem cell propagation and chemoresistance.
        J Biol Chem. 2013; 288: 4334-4345
        • Mione M.
        • Bosserhoff A.
        MicroRNAs in melanocyte and melanoma biology.
        Pigment Cell Melanoma Res. 2015; 28: 340-354
        • Muller J.
        • Krijgsman O.
        • Tsoi J.
        • Robert L.
        • Hugo W.
        • Song C.
        • et al.
        Low MITF/AXL ratio predicts early resistance to multiple targeted drugs in melanoma.
        Nat Commun. 2014; 5: 5712
        • Muramatsu F.
        • Kidoya H.
        • Naito H.
        • Sakimoto S.
        • Takakura N.
        microRNA-125b inhibits tube formation of blood vessels through translational suppression of VE-cadherin.
        Oncogene. 2013; 32: 414-421
        • Nyholm A.M.
        • Lerche C.M.
        • Manfe V.
        • Biskup E.
        • Johansen P.
        • Morling N.
        • et al.
        miR-125b induces cellular senescence in malignant melanoma.
        BMC Dermatol. 2014; 14: 8
        • Queirolo P.
        • Picasso V.
        • Spagnolo F.
        Combined BRAF and MEK inhibition for the treatment of BRAF-mutated metastatic melanoma.
        Cancer Treat Rev. 2015; 41: 519-526
        • Rambow F.
        • Job B.
        • Petit V.
        • Gesbert F.
        • Delmas V.
        • Seberg H.
        • et al.
        New functional signatures for understanding melanoma biology from tumor cell lineage-specific analysis.
        Cell Rep. 2015; 13: 840-853
        • Ritchie M.E.
        • Phipson B.
        • Wu D.
        • Hu Y.
        • Law C.W.
        • Shi W.
        • et al.
        limma powers differential expression analyses for RNA-sequencing and microarray studies.
        Nucleic Acids Res. 2015; 43: e47
        • Roesch A.
        • Fukunaga-Kalabis M.
        • Schmidt E.C.
        • Zabierowski S.E.
        • Brafford P.A.
        • Vultur A.
        • et al.
        A temporarily distinct subpopulation of slow-cycling melanoma cells is required for continuous tumor growth.
        Cell. 2010; 141: 583-594
        • Sanz-Moreno V.
        • Gadea G.
        • Ahn J.
        • Paterson H.
        • Marra P.
        • Pinner S.
        • et al.
        Rac activation and inactivation control plasticity of tumor cell movement.
        Cell. 2008; 135: 510-523
        • Schlegel N.C.
        • von Planta A.
        • Widmer D.S.
        • Dummer R.
        • Christofori G.
        PI3K signalling is required for a TGFbeta-induced epithelial-mesenchymal-like transition (EMT-like) in human melanoma cells.
        Exp Dermatol. 2015; 24: 22-28
        • Singh M.K.
        • Izumchenko E.
        • Klein-Szanto A.J.
        • Egleston B.L.
        • Wolfson M.
        • Golemis E.A.
        Enhanced genetic instability and dasatinib sensitivity in mammary tumor cells lacking NEDD9.
        Cancer Res. 2010; 70: 8907-8916
        • Sobrado V.R.
        • Moreno-Bueno G.
        • Cubillo E.
        • Holt L.J.
        • Nieto M.A.
        • Portillo F.
        • et al.
        The class I bHLH factors E2-2A and E2-2B regulate EMT.
        J Cell Sci. 2009; 122: 1014-1024
        • Sun Y.M.
        • Lin K.Y.
        • Chen Y.Q.
        Diverse functions of miR-125 family in different cell contexts.
        J Hematol Oncol. 2013; 6: 6
        • Ugurel S.
        • Rohmel J.
        • Ascierto P.A.
        • Flaherty K.T.
        • Grob J.J.
        • Hauschild A.
        • et al.
        Survival of patients with advanced metastatic melanoma: the impact of novel therapies.
        Eur J Cancer. 2015; 53: 125-134
        • Verfaillie A.
        • Imrichova H.
        • Atak Z.K.
        • Dewaele M.
        • Rambow F.
        • Hulselmans G.
        • et al.
        Decoding the regulatory landscape of melanoma reveals TEADS as regulators of the invasive cell state.
        Nature Commun. 2015; 6: 6683
        • Vergani E.
        • Di Guardo L.
        • Dugo M.
        • Rigoletto S.
        • Tragni G.
        • Ruggeri R.
        • et al.
        Overcoming melanoma resistance to vemurafenib by targeting CCL2-induced miR-34a, miR-100 and miR-125b.
        Oncotarget. 2015; 7
        • Widmer D.S.
        • Cheng P.F.
        • Eichhoff O.M.
        • Belloni B.C.
        • Zipser M.C.
        • Schlegel N.C.
        • et al.
        Systematic classification of melanoma cells by phenotype-specific gene expression mapping.
        Pigment Cell Melanoma Res. 2012; 25: 343-353
        • Zhang L.
        • Ge Y.
        • Fuchs E.
        miR-125b can enhance skin tumor initiation and promote malignant progression by repressing differentiation and prolonging cell survival.
        Genes Dev. 2014; 28: 2532-2546
        • Zhang L.
        • Stokes N.
        • Polak L.
        • Fuchs E.
        Specific microRNAs are preferentially expressed by skin stem cells to balance self-renewal and early lineage commitment.
        Cell Stem Cell. 2011; 8: 294-308