Demethylation by low-dose 5-aza-2′-deoxycytidine impairs 3D melanoma invasion partially through miR-199a-3p expression revealing the role of this miR in melanoma
- 289 Downloads
Efficient treatments against metastatic melanoma dissemination are still lacking. Here, we report that low-cytotoxic concentrations of 5-aza-2′-deoxycytidine, a DNA demethylating agent, prevent in vitro 3D invasiveness of metastatic melanoma cells and reduce lung metastasis formation in vivo.
We unravelled that this beneficial effect is in part due to MIR-199A2 re-expression by promoter demethylation. Alone, this miR showed an anti-invasive and anti-metastatic effect. Throughout integration of micro-RNA target prediction databases with transcriptomic analysis after 5-aza-2′-deoxycytidine treatments, we found that miR-199a-3p downregulates set of genes significantly involved in invasion/migration processes. In addition, analysis of data from melanoma patients showed a stage- and tissue type-dependent modulation of MIR-199A2 expression by DNA methylation.
Thus, our data suggest that epigenetic- and/or miR-based therapeutic strategies can be relevant to limit metastatic dissemination of melanoma.
KeywordsDNA methylation microRNA Melanoma aggressiveness Epigenetics
CpG island methylator phenotype
DNA methyltransferase inhibitors
Distant organ metastases
Half effective concentration
Lymph nodes metastases
Regional cutaneous tissues
Skin cutaneous melanoma
The Cancer Genome Atlas
Cutaneous metastatic melanoma is the deadliest form of skin cancer, with a 5-year survival rate particularly low (< 10%) and increasing incidence . Until recently, patients with metastatic melanoma had few treatment options, including surgery, radiation therapy and/or chemotherapy, giving transient and limited results. Since 2011, new targeted therapies against mutated BRAF (vemurafenib) and immunotherapies such as anti-CTLA4 (ipilimumab) and anti-PD1/PD-L1 antibodies emerged, giving very promising long-term responses. Despite the efficacy of these treatments in clinical trials, resistance rapidly arises after vemurafenib and immunotherapies can induce severe side effects limiting the therapeutic response . In addition, not all patients respond to these therapies and no molecular marker is available to predict patient responsiveness to immunomodulators. Therefore, there is still a strong need for novel strategies.
Interestingly, targeting of the epigenetic regulation is promising for the development of new anticancer treatments in solid tumours . In the past few years, various studies reported on epigenetic alterations in metastatic melanoma, more specifically DNA methylation, and associated them with disease progression [4, 5, 6, 7, 8, 9]. DNA methylation, catalysed by DNA methyltransferases (DNMT), is a well-studied epigenetic change in cancer. In the last decade, several epigenetic drugs, such as DNA methylation inhibitors (DNMTi), azacitidine (5aza, Vidaza®) and decitabine (5azadC, Dacogen™), have been approved for the treatment of haematological tumours. In solid tumours, the use of epigenetic therapies is breaking through . Hypomethylation of repetitive DNA elements and hypermethylation of CpG islands located in promoter regions of specific genes are hallmarks of melanoma malignancy [10, 11, 12, 13]. Specific methylation signatures were also found for melanoma tumours harbouring BRAF mutations [8, 14]. Thus, the use of an epigenetic strategy, combined with relevant CpG island methylator phenotype (CIMP)  as markers for the prediction of prognosis or therapeutic response, could be envisioned for metastatic melanoma treatment [10, 16, 17].
In this context, we investigated the implication of DNA methylation in the aggressiveness of metastatic melanoma and whether the use of demethylating agents can reverse its invasive phenotype. We focused our attention on microRNAs (miRs), since these endogenous non-coding RNAs of 19–25 nt can modulate the expression of several target genes via translational blockade and/or transcript degradation and are broadly involved in cancer progression. Furthermore, some miRs are described to be epigenetically regulated [12, 18] and to participate in the acquisition of invasive capabilities of cancer cells, including melanoma [19, 20, 21]. Here, we identified one specific hypermethylated miR gene, MIR199A2, whose expression restoration by low doses of 5-aza-2′-deoxycytidine or miR transfection impacts on in vitro cell invasiveness and in vivo metastatic formation.
5azadC impairs melanoma cell invasiveness in a 3D invasion model at low non-cytotoxic doses
Hence, low concentrations of 5azadC, inducing little cell death and inhibition of cell proliferation (EC50 = 100 nM), were chosen for the 3D invasion assay to limit non-specific cytotoxic effects and favour an epigenetic effect. As illustrated in Fig. 1b, c, 5azadC induced a dose-dependent inhibition of 3D cell invasion, with a significant decrease of the invasion index starting at 3.2 nM (P value < 0.001). The anti-metabolic compound cytarabine (araC), structurally similar to 5azadC and commonly used in chemotherapies , was used as control at concentrations resulting in < 10% of cell death (1 nM, Additional file 1: Figure S3A). Unlike 5azadC, at the equi-cytotoxic concentration, araC did not cause cell invasion inhibition (Additional file 1: Figure S3B and S3C), confirming a different mechanism of action for 5azadC.
Next, the DNA demethylating activity of the drug was determined by following the global DNA methylation level in spheroids recovered before inclusion in collagen (at day 7, Fig. 1a). The methylation of four CpG sites in LINE-1 elements was chosen as a surrogate marker for global genomic DNA methylation . Figure 1d shows that LINE-1 methylation decreased significantly upon treatment with 5azadC concentrations as low as 1 nM (P value < 0.05). This demethylating effect was dose-dependent up to the EC50 measured at day 7 (100 nM).
Altogether, our data revealed that 5azadC displays an anti-invasive effect in a 3D metastatic melanoma model at low concentrations. This effect is correlated to its DNA demethylating action but not to its anti-metabolic and cytotoxic properties, as deduced from lack of effect of araC.
DNA methylation modulation by 5azadC reactivates miR-199a-3p
To shed light on the mechanisms involved in the 5azadC-associated anti-invasive effect, we searched for genes and in particular miRs that were hypermethylated in the metastatic cell line WM-266-4 and which expression could be restored by promoter demethylation. Using data from a genome-wide DNA methylation analysis (BeadChip Illumina 450 K, data not shown), we compared the methylation profile of WM-266-4 (metastatic cells) and its non-invasive counterpart WM-115 cell lines (derived from the primary tumour of the same patient). All genes and miR hypermethylated in WM-266-4 cell line with a difference of at least 20% on three CpGs were selected. Among the 68 hypermethylated miRs, two were chosen for further analysis since they were previously described as playing a role in cell invasion/migration processes: MIR-155HG  and MIR-199A2 [30, 31].
The methylation status of CpG sites in the promoter of these two miRs was examined after 5azadC treatment by DNA bisulfite conversion followed by pyrosequencing. For MIR-155HG, CpG sites were found hypermethylated in the non-treated WM-266-4 GFP spheroids with a median methylation of 88% (Fig. 1e). Upon 5azadC treatment, the decrease of methylation was comprised between 10 and 23%. For the MIR-199A2 promoter, CpG sites were found methylated at 73% in non-treated WM-266-4 GFP spheroids and started to be demethylated from 1 nM of 5azadC. A maximum demethylation of 31% was reached at 100 nM (Fig. 1e).
To correlate this DNA demethylation with the respective RNA expression, mature miR levels were monitored in parallel. A dose-response increase of miR-155 expression was found upon 5azadC treatment without statistical significance (Fig. 1f). Regarding miR-199a, two mature miR species are known to be derived from the miR-199a precursor: miR-199a-5p and miR-199a-3p, the former being 13 times less expressed than the latter (Fig. 1f, RQ to RNU6B of 0.086 ± 0.023 vs. 1.16 ± 0.14). A significant RNA re-expression of nearly threefold was found with 100 nM of 5azadC for miR-199a-3p, while no significant difference was observed for miR-199a-5p (Fig. 1f).
Hence, these results showed that low concentrations of 5azadC lead to DNA demethylation of specific hypermethylated miR promoters in metastatic melanoma WM-266-4 GFP cells. More specifically, MIR-199A2 promoter demethylation was correlated with the re-expression of its mature form miR-199a-3p.
MiR-199a-3p impairs melanoma cell invasiveness and is involved in the anti-invasive effect of 5azadC
To confirm this hypothesis, miR-199a-3p expression was knocked down in 5azadC-treated cells using a specific anti-miR-199a-3p (Fig. 2b and Additional file 1: Figure S4B). The 3D anti-invasive ability of 5azadC was significantly reduced, and the invasion index was partially restored (Fig. 2b). These findings support our hypothesis that miR-199a-3p plays a role in the 5azadC-mediated anti-invasive effect.
MET targeting by miR-199a-3p is not sufficient to affect 3D cell invasion
These results suggest an inverse relationship between miR-199a-3p and MET and thus a possible direct targeting of this oncogene by the miR. A similar decrease of MET was also observed upon 5azadC treatment (Fig. 3c), prompting us to hypothesise that re-expression of miR-199a-3p following DNA demethylation participates in the downregulation of MET. Nevertheless, MET knock down was not sufficient to inhibit 3D invasiveness of metastatic melanoma as shown in Fig. 3d. Therefore, MET depletion seemed not to be the major event leading to miR-199a-3p or 5azadC anti-invasive effect in our model.
MiR-199a-3p and 5azadC deregulate common genes involved in invasion/migration pathways
Both 5azadC and miR-199a-3p alone impair lung metastasis formation by WM-266-4 cells in vivo
Promoter methylation status and expression of MIR-199A2 in patient samples
Taken together, these data support our hypothesis that MIR-199A2 hypermethylation can be important at an early stage for the acquisition of an invasive phenotype by the melanoma tumour cells.
Epigenetic therapies are very promising strategies to fight cancer, since epigenetic modifications are reversible and can be chemically modulated. In particular, promoters’ demethylation by 5-azaC or 5-azadC induces the re-expression of several genes or miRs in cancer cells inducing an anti-cancer effect . To date, few clinical trials using epigenetic drugs have been conducted in metastatic melanoma; a recent example is the combination of 5azadC and temozolomide [38, 39]. Yet, epigenetic modifications and, in particular, genomic hypermethylation of CpG islands and miR modulation were shown to sign metastatic melanoma progression [40, 41].
To the best of our knowledge, few reports explored the impact of 5azadC on the cellular mechanism involved in metastatic melanoma, particularly on cell invasiveness, and its potential use as an anti-metastatic drug . To address this question and better mimic the tumour behaviour in vivo, we developed a 3D cell invasion assay using spheroids embedded in a matrix of collagen I, the principal component of the extracellular matrix (ECM). Growing evidence showed that cancer cell spheroids reproduce various traits of micro-tumours, thus providing a better model for assessing tumour invasion . Using the metastatic melanoma WM-266-4 GFP cell line, we showed for the first time that low-cytotoxic concentrations of 5azadC inhibit 3D cell invasion in vitro. In addition, this effect was clearly associated with a global and promoter-specific DNA demethylation. These observations are in complete agreement with Rajaii et al.’s results on uveal melanoma cells, describing that low concentrations of 5azadC reduced both 2D invasion and clonogenicity . Our results in metastatic melanoma cells confirm also the anti-cancer potential of using low doses of the DNA demethylating agents in solid tumours as for haematological cancers [27, 38, 42].
To further elucidate the mechanisms underlying the phenotypical changes induced by 5azadC, we specifically investigated the modifications on miR methylation and expression. MiRs are post-transcriptional regulators that have a broad spectrum of cellular regulation by targeting multiple genes and pathways, including those required for metastatic melanoma progression like cell invasion and migration (reviews ). Moreover, their expression or DNA methylation signatures are studied as reliable markers for diagnosis and prognosis of cancer . Here, two miRs, MIR-155HG and MIR-199A2, implicated in migration and invasion processes, were found hypermethylated in the metastatic melanoma cell line (WM-266-4) compared to its primary counterpart (WM-115). MiR-155 is an important multi-functional miR involved in the development of multiple solid tumours and is transcriptionally regulated by DNA methylation . However, 5azadC-induced demethylation of MIR-155HG promoter (> 20%) was not sufficient to significantly reactivate miR-155 expression in our metastatic melanoma cell model. Moreover, its ectopic re-expression in the WM-266-4 cells induced a cytotoxic effect rather than an anti-invasive effect. Thus, we focused on MIR-199A2 that was found hypermethylated in the WM-266-4 cell line and both demethylated and reactivated by 5azadC.
Few reports identified epigenetically modulated miRs in melanomas (reviews [40, 41]). For the first time, we report that MIR-199A2 locus is under epigenetic regulation in a metastatic melanoma model and it is reactivated by promoter demethylation by 5azadC. In addition, nanomolar concentrations of miR-199a-3p impairs 3D cell invasion, unlike miR-199a-5p. Furthermore, the anti-invasive effect of 5azadC was counterbalanced by a specific antago-miR suggesting that miR-199a-3p takes part of the anti-invasive response induced by 5azadC. Importantly, the transient expression of miR-199a-3p in WM-266-4 cells reduced significantly their metastatic potency in vivo, supporting an anti-cancer effect of this miR.
These results are somewhat contradictory to those reported by Pencheva et al. (2012) , describing that the cooperation between three miRs, miR-199a-3p, miR-199a-5p and miR-1908, promoted metastasis by impairing two proteins, APOE and DNAJA4, involved in tumour cell invasion and in angiogenesis. This discrepancy can be explained as follows. First, the cellular models used are significantly different. Pencheva et al. selected for their study the most in vivo metastatic subclones of MeWo and A375, two metastatic melanoma cell lines bearing BRAF wild-type and BRAFV600E mutation, respectively. These subclones overexpressed miR-199a-3p and miR-199a-5p in different proportions. The A375 subclones, bearing a mutated BRAF, as it is the case of the WM266-4 cell line, overexpressed much higher levels of miR-199a-5p (× 10,000) than of miR-199a-3p (× 10), strongly suggesting that mainly the miR-199a-5p supported the pro-metastatic effect that was observed by the authors in these subclones. In agreement with this hypothesis, we observed a reversion of the anti-invasive effect of miR-199a-3p when combined with increasing amounts of the complementary miR-199a-5p mimetic (data not shown), revealing a sponge effect of the latter. Second, Pencheva et al. performed in vivo experiments with the combination of the three LNA anti-miRs, targeting miR-199a-3p, miR-199a-5p and miR-1908 but not with individual ones or the miR-199a-3p mimetic alone. Thus, it cannot be excluded that the balance between each miR level can affect the tumour development. Moreover, cumulative evidences have shown that miR-199a has important but versatile functions in tumourigenesis depending on the nature and the stage of the cancer . Indeed, Yang et al. found that low expression levels of miR-199a-5p in tumour melanoma tissue samples from patients were associated to the advances of tumour stage . Their observations in B16 melanoma cells and in tumour xenografts seem to indicate that miR-199a-5p is a tumour suppressor, via direct targeting of HIF-1α. Third, our findings in metastatic melanoma are in agreement with an increasing number of observations in other cancer types. In testicular tumour, MIR-199A2 hypermethylation was linked to malignancy progression . In this model, miR-199a re-expression led to the suppression of cell growth, cancer migration, invasion and metastasis in vitro and in vivo. Similar observations were made in non-small cell lung cancer, colorectal cancer and breast cancer cell lines . A tumour-suppressive role of this miR was also demonstrated in papillary thyroid carcinoma, in which miR-199a-3p was shown to be likely regulated by the BRAFV600E oncogene . Similar anti-cancer functions were described in renal cell carcinoma , endometrial carcinoma  and prostate cancer .
Finally, to better understand the mechanisms involved in the miR-199a-3p-mediated anti-invasive effect in metastatic melanoma, its target genes were investigated. The proto-oncogene MET was found overexpressed in the metastatic vs. primary cell line. Despite its downregulation by both miR-199a-3p and 5azadC, its specific inhibition was not sufficient to impair 3D metastatic melanoma invasion (Fig. 4). Next, by a transcriptomic and bioinformatic approach, we observed that miR-199a-3p targeted several genes involved in key pathways for cell migration/invasion. This alteration of multiple targets can better explain the in vitro anti-invasive and in vivo anti-metastatic effects of the miR. Interestingly, 5azadC also downregulated common effectors of this multistep biological process (Fig. 5). In addition, analysis of TCGA SKCM datasets supports a tight modulation of MIR-199A2 promoter methylation and miR-199a expression during melanoma progression, with local stages (RCT) presenting a promoter demethylation and an increased expression, whereas disseminated tumours (LN) show a MIR-199A2 hypermethylation concomitant to a decreased repression (Fig. 7). These observations were further confirmed on an experimental cohort of primary melanoma, lymph node and distant metastasis. Altogether, these observations can explain the different effects found in the literature for miR-199a-3p depending on its targets and strongly promote the interest of its reactivation at early melanoma stages that start to locally disseminate.
In conclusion, we proved that slightly cytotoxic doses of 5azadC are translated into an anti-invasive effect in a BRAF-mutated metastatic melanoma model. This effect is in part associated with MIR-199A2 promoter demethylation and re-expression of its specific mature form miR-199a-3p. Indeed, this miR on its own impairs 3D cell invasion by simultaneously targeting multiple genes involved in adhesion/migration processes. This effect was further confirmed in an in vivo xenograft model with transiently transfected cells, thus supporting its implication in metastatic melanoma aggressiveness. Finally, the clinical relevance of our results is confirmed by the analysis of the TCGA cancer skin cohort and 43 patient samples, showing that MIR-199A2 promoter methylation and expression is finely tuned during melanoma progression. Hypermethylation of its promoter was found associated with LN metastasis stages before distant dissemination (Fig. 7), suggesting its expression extinction could be involved in metastasis progression. Taken together, our findings support the use of epigenetic drugs or miR mimetic-based therapeutic strategies, either alone or in combination, for limiting metastatic dissemination of melanomas. Concordantly, a combination of 5azadC with temozolomide was recently investigated in a phase I/II trial and proven to be safe and improve response and overall survival rates .
Materials and methods
The details of the protocols, including primer sequences, are reported in Additional file 1.
Human cell lines
The WM-115 and WM-266-4 cell lines were purchased from the European Collection of Authenticated Cell Culture (ECACC), were subjected to cell line authentication via short tandem repeat profiling and were grown up to a maximum of 20 passages and for fewer than 6 months following resuscitation in our laboratory. The WM-266-4 GFP cell line was obtained by transduction with WPXLd, a lentiviral vector encoding GFP, and selected by flow cytometry. Additionally, four primary melanoma cell lines (WC00060, WC00062, WC0008 and WC00081) were purchased from the Coriell Cell Repository (Coriell Institute for Medical Research, Camden, NJ, USA).
5-Aza-2′-deoxycytidine (5azadC, decitabine) and cytarabine (araC) from Sigma-Aldrich (St Louis, USA) were dissolved in water at 10 mM and frozen at − 20 °C until use. Serial dilutions were made in culture media and added on cells seeded 24 h earlier at a density of 0.6 × 105 cells/mL in a 12-well plate. Treatment was repeated during 3 days, and cells were harvested at day 4 for further analysis.
Cell viability assay
The half effective concentration (EC50) of 5azadC or araC was determined at day 7 after the first treatment on WM-266-4 GFP cells using ATP-lite™ assay (Perkin Elmer, France) according to the manufacturer for cells cultured in 2D conditions. This protocol was also adapted for spheroids (3D conditions).
3D cell invasion assay
WM-266-4 GFP (3000 cells/well) were allowed to form spheroids for 2 days on agarose 1% (Sigma-Aldrich, #A95–39) coated in 96-well plates, leading to ~ 300-μm-diameter spheroids. Then, six different spheroids for each condition were individually embedded in EMEM media (Lonza, #BE12-684F) containing 1% of Bovine Collagen I (BD, #354231) and 2% SVF. The initial spheroid size (day 7) and the 24-h invasion area (day 8) were measured by fluorescent microscopy, using an Axiovert 200 M device (5X Plan-Neofluor objective, Carl Zeiss, Germany). Fluorescent invasion areas were quantified using ImageJ (NIH) software on the sum of six Z-stacks images (20-μm interval) for each spheroid. Invasion index was obtained by normalising the invasion area at 24 h by the initial spheroid area: invasion area/spheroid area (Additional file 1: Figure S1). Six individual spheroids were quantified for each condition, and three independent experiments were performed.
Genomic DNA purification
Paraffin-embedded tissue sections were microdissected, and DNA was purified using the ZR FFPE DNA MiniPrep (Zymo Research, Irvine, CA) as reported . Genomic DNA from melanoma cell lines was extracted using DNAzol reagent (Life Technologies, Carlsbad, CA) as in .
Bisulfite modification and pyrosequencing
Quantitative DNA methylation analysis was performed by pyrosequencing of bisulfite-treated DNA as described in . Two PCR were carried out for the five CpGs of interest on the miR-199a-3p locus.
PCR1: (1 CpG):
Pyrosequencing primer: GGTTGTGATTTTTAGTTTTGAYGTGGT
PCR2: (4 CpGs):
Microarray expression and GO terms analysis
RNA samples were extracted from cells transfected with miR-199a-3p miR mimetic as described in Additional file 1. Experiments were performed in three biological replicates for each condition and analysed on Affymetrix Human Gene 2.0 ST arrays. RMA normalisation was applied, and genes differentially expressed between treated and control conditions were determined choosing a threshold of log2 fold change at 0.5 and a P value < 0.05. P value cutoff is FDR corrected . Significantly enriched pathways were chosen using a threshold enrichment score > 2. Data have been submitted to the Gene Expression Omnibus database.
In vivo metastasis experiments
The animals were handled and cared for in accordance with the Guide for the Care and Use of Laboratory Animals (National Research Council, 1996) and European Directive EEC/86/609, under the supervision of the authorised investigators. Un-anesthetised 7-week-old female SCID mice (Charles River Laboratories, Saint-Germain-sur-l’Arbresle, France) were injected into the tail vein with 3 × 106 viable WM-266-4 cells in 200-μL PBS, previously treated or not with 5azadC at 100 nM (3 × 3 days) or transfected with miR negative control #1 (miR Ctl) or miR-199a-3p mimic at 0.5 or 1 nM during 48 h. Each group constituted of 4–5 mice. For 5azadC treatment, two independent experiments were performed. Twenty-one days after injection, mice were dissected and lungs recovered, formalin fixed and paraffin embedded. Sections were stained with haematoxylin and eosin (H&E). The number and area of macrometastasis were measured in total lung sections by immunostaining with Tyrosinase antibody Mob299–05 (1/500) (Diagnostic BioSystem, Pleasanton, CA-USA) on an Aperio Scanner (Leica) with Tissue Studio software (Definiens, Munich, Switzerland) for automatic segmentation and quantification of lesions. Statistics were performed using the Mann-Whitney test.
The Cancer Genome Atlas skin cutaneous melanoma (TCGA-SKCM) dataset analysis
The TCGA-SKCM DNA methylation, RNA-sequencing (RNAseq) and clinical datasets were accessed in June 2016. Wilcoxon-Mann-Whitney test was applied to determine the significance of DNA methylation or gene expression differences between primary and metastatic melanoma tissues.
Melanoma patients included in the study
Melanoma patients and healthy individuals were included in the study under protocols approved by the Western institutional review board. Informed consent was obtained from all subjects, and the experiments were performed according to the principles set out in the WMA Declaration of Helsinki and the NIH Belmont Report. Tissue specimens were coded according to HIPAA recommendations to ensure the confidentiality of the patients. The specimen cohort included paraffin-embedded tissues including nevi (NEV; n = 7), primary melanomas (PRM; n = 16), lymph node metastases (LNM; n = 15) and distant organ metastases (DOM; n = 20). Additionally, cell lines derived from LNM (n = 11) were established at JWCI and included in the study. For each patient, median DNA methylation was calculated on five CpGs on the miR-199a-3p locus.
This work was supported to P.B.A. by Centre National de la Recherche Scientifique (CNRS) [ATIP], Région Midi Pyrenées [Equipe d’Excellence and FEDER CNRS/Région Midi Pyrenées], Fondation InNaBioSanté, and PlanCancer2014 (N°EPIG201401).
Availability of data and materials
In addition to the data available in the supplementary information files, datasets used and/or analysed during the current study are available from the corresponding author. Materials are also available on reasonable request.
CD conceived and performed the experiments and wrote the manuscript. AC characterised the DNA methylation, analysed the patient data and performed the experiments. AD set up the 3D invasion model. AD, FB and JT performed the pyrosequencing assays. AP and JR performed the in vivo experiments. DM contributed with the data analysis and DNA purification from melanoma tissues and cell lines. JR and GF participated in the design of the study. CE and PBA oversaw and designed the study and wrote the manuscript. All authors read and approved the final manuscript.
Ethics approval and consent to participate
Informed consent was obtained from all subjects, and experiments were performed according to the principles set out in the WMA Declaration of Helsinki and the NIH Belmont Report. Tissue specimens were coded according to HIPAA recommendations to ensure the confidentiality of the patients.
Consent for publication
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
- 10.Besaratinia A, Tommasi S. Epigenetics of human melanoma: promises and challenges. J Mol Cell Biol. 2014.Google Scholar
- 35.Bachmann IM, Halvorsen OJ, Collett K, Stefansson IM, Straume O, Haukaas SA, et al. EZH2 expression is associated with high proliferation rate and aggressive tumor subgroups in cutaneous melanoma and cancers of the endometrium, prostate, and breast. J Clin Oncol. 2006;24(2):268–73.PubMedCrossRefGoogle Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.