Plasticity of melanoma and EMT-TF reprogramming

For several decades, tumour progression has been considered as linear process involving clonal selection of most aggressive cell variants. Within the framework of this concept, multiple attempts have been made to identify genes specifically activated at late cancer stages and involved in cancer metastases. However, there are no experimental evidences supporting the existence of such purely “metastatic genes”, and metastatic behaviour is likely to be induced by the same mutationally-activated pathways that cause neoplastic transformation. In line with this view is the parallel model of tumour progression and metastases. It proposes that at least in some cancer types early metastatic spread is concomitant with oncogenic transformation, i.e. the same genetic insults are responsible for the tumourigenicity and metastatic dissemination [1, 2]. Epithelial-mesenchymal transition (EMT) is a reversible embryonic genetic program reactivated in cancer. EMT is controlled by several transcription factors, EMT-TFs (TWIST1, TWIST2, SNAIL1, SNAIL2, ZEB1, ZEB2, etc.). The initial view on EMT as just a process leading to enhanced cell motility and invasion was reconsidered in recent years. Aberrant activation of EMT programs in tumour cells affects cell proliferation, survival, drug resistance, tumourigenicity and stem cell-like features [3]. In particular, a proinvasive EMT-TF, TWIST1 cooperates with activated oncogenes by overriding oncogene-induced senescence [2]. These findings propose that EMT may act in concert with classical oncogenic pathways and contribute to different stages of tumourigenesis including tumor initiation, growth and spread. We addressed the interrelationship between EMTTF network and tumour-initiating pathways in malignant melanoma [4]. Melanoma represents an example of a cancer type solely dependent on the oncogenic pathways initiated by gain-of function mutations in BRAF, NRAS, GRM3 or MEK1/2 with all of them leading to the activation of MEK-ERK module. Melanocytic lineage is distinct from epithelium; it evolves from neural crest, an embryonic cell population regulated by EMT pathways. Our data show that normal melanocytes are positive for SNAIL2 and ZEB2, but negative for ZEB1 and TWIST1. In vitro activation of the MEK-ERK signalling results in the reversion of EMT-TF expression pattern, the downregulation of SNAIL2 and ZEB2 and upregulation of ZEB1/TWIST1. This reversible EMTTF reprogramming is followed by the reduction in the expression of a master regulator of melanocytic lineage, microphthalmia-associated transcription factor (MITF), and repression of the downstream differentiation program. Moreover, switch to ZEB1 and TWIST1 is utterly required for the tumourigenic potential of BRAF in vitro and in vivo and for the increased invasiveness of BRAFtransformed melanocytes. Though the absolute majority of melanocytic lesions contain MEK-activating mutations, they exhibit a large degree of heterogeneity with regard to phospho-ERK immunopositivity. This is largely due to the activation of different negative feedback pathways in benign nevi and horizontal phase melanoma, and acquisition of bypass mechanisms in advanced cancer. Our analyses of melanoma samples have shown that EMT-TF reprogramming does exist in human melanoma. It strongly correlated with the phospho-ERK immunopositivity and poor patients’ survival. Gradients of EMT-TF expression were observed in both the primary melanoma and independent or matched metastases. Namely, expression of ZEB2/SNAIL2 detected in superficial parts of primary tumours and cortical areas of lymph node metastases gradually decreased with the distance from the edge of the lesion. The expression of ZEB1 and TWIST1 was observed in deeper parts of primary tumours and medullary areas of nodal metastases and significantly correlated with phosphoERK immunopositivity. The reestablishment of EMT-TFs gradients at secondary sites may indicate the existence Editorial Material