Understanding Molecular Mechanisms of Phenotype Switching and Crosstalk with TME to Reveal New Vulnerabilities of Melanoma

Abstract

Melanoma cells are notorious for their high plasticity and ability to switch back and forth between various melanoma cell states, enabling the adaptation to sub-optimal conditions and therapeutics. This phenotypic plasticity, which has gained more attention in cancer research, is proposed as a new paradigm for melanoma progression. In this review, we provide a detailed and deep comprehensive recapitulation of the complex spectrum of phenotype switching in melanoma, the key regulator factors, the various and new melanoma states, and corresponding signatures. We also present an extensive description of the role of epigenetic modifications (chromatin remodeling, methylation, and activities of long non-coding RNAs/miRNAs) and metabolic rewiring in the dynamic switch. Furthermore, we elucidate the main role of the crosstalk between the tumor microenvironment (TME) and oxidative stress in the regulation of the phenotype switching. Finally, we discuss in detail several rational therapeutic approaches, such as exploiting phenotype-specific and metabolic vulnerabilities and targeting components and signals of the TME, to improve the response of melanoma patients to treatments.

Keywords: melanoma; metabolic reprogramming; oxidative stress; phenotype switching; therapeutic strategies; tumor microenvironment.

Figure 1

Schematic description of the complex spectrum of phenotype switching in melanoma. The distinct melanoma states and their phenotype specific markers, the epigenetic characteristics, the metabolic reprogramming, and the oxidative stress status among the main melanoma phenotypes.

Figure 2

Melanoma cell states associated with minimal residual disease (MRD). Exposure to MAPK inhibitors promotes a phenotypic switch from the melanocytic state towards a starved-like state (SMC) facilitating a transition towards either a pigmented/differentiated phenotype or to an undifferentiated one: the classical invasive state or the neural crest stem-like cell (NCSC) state (adapted from Rambow et al. 2018 [43]).

Figure 3

Comparison of melanoma signatures across different publications. Venn diagram (http://bioinformatics.psb.ugent.be/webtools/Venn/) showing the distribution of shared and special genes of main melanoma states across different publications ([5,7,25,43]): (A) the melanocytic (proliferative/differentiated) signatures (Hoek proliferative (Pro) [7], Verfaillie proliferative (Pro) [5], Tsoi melanocytic [25]), and a list of common genes among these signatures. (B) The mesenchymal-like (invasive/undifferentiated) signatures (Hoek invasive (Inv) [7], Verfaillie invasive (Inv) [5], Tsoi undifferentiated [25], Rambow invasive (Inv) [43]), and the list of common genes among these signatures.

Figure 4

Phenotype switching in melanoma: prominent role of microenvironment stress signals and oxidative stress. Several microenvironmental cues, such as hypoxia, inflammation, and nutrient deprivation, can promote oxidative stress leading to eIF2α phosphorylation and eIF2B inhibition and, consequently, to the synthesis of adaptive stress proteins, including ATF4, and to the inhibition of global protein translation, which repress MITF. Translation reprogramming and MITF repression can induce a phenotypic switch towards an invasive state.

Figure 5

ROS generation in melanoma cells promoting invasive phenotype. ROS can be produced by melanosomes, the NOX family, and mitochondria. High ROS levels can promote the invasive phenotype.

Figure 6

MITF-PGC1α axis in melanoma plasticity. High correlation between MITF and PGC1α expression across melanoma phenotypes in the panel of 53 human melanoma lines used in Tsoi et al. 2018 [25]. PCA of melanoma cell line expression profiles are annotated by identified clusters. MITF and PGC1α are highly expressed in the melanocytic (C4-green) and the transitory (C3-purple) phenotypes whereas this expression is low in the invasive phenotypes (neural crest-like (red) and undifferentiated (blue) phenotypes). Illustration was created by using the interactive web interface resource [25] available at http://systems.crump.ucla.edu/dediff/.

Figure 7

Targeted therapeutic approaches based on melanoma states vulnerabilities, metabolic rewiring, and microenvironment modulation. Schematic description of different therapeutic strategies: (A) exploiting vulnerabilities of distinct phenotypes, (B) targeting lipid metabolism, and (C) modulating components and signals of tumor microenvironment (TME).