BackgroundDuring embryonic development Wnt family members and bone morphogenetic proteins (BMPs) cooperatively induce epithelial-mesenchymal transition (EMT) in the neural crest. Wnt and BMPs are reactivated during malignant transformation in melanoma. We previously demonstrated that the BMP-antagonist noggin blocked the EMT phenotype of melanoma cells in the neural crest and malignant invasion of melanoma cells in the chick embryo; vice-versa, malignant invasion was induced in human melanocytes in vivo by pre-treatment with BMP-2.ResultsAlthough there are conflicting results in the literature about the role of β-catenin for invasion of melanoma cells, we found Wnt/β-catenin signaling to be analogously important for the EMT-like phenotype of human metastatic melanoma cells in the neural crest and during invasion: β-catenin was frequently expressed at the invasive front of human primary melanomas and Wnt3a expression was inversely correlated with survival of melanoma patients. Accordingly, cytoplasmic β-catenin levels were increased during invasion of melanoma cells in the rhombencephalon of the chick embryo. Fibroblast derived Wnt3a reduced melanoma cell adhesion and enhanced migration, while the β-catenin inhibitor PKF115–584 increased adhesion and reduced migration in vitro and in the chick embryonic neural crest environment in vivo. Similarly, knockdown of β-catenin impaired intradermal melanoma cell invasion and PKF115–584 efficiently reduced liver metastasis in a chick chorioallantoic membrane model. Our observations were accompanied by specific alterations in gene expression which are linked to overall survival of melanoma patients.ConclusionWe present a novel role for Wnt-signaling in neural crest like melanoma cell invasion and metastasis, stressing the crucial role of embryonic EMT-inducing neural crest signaling for the spreading of malignant melanoma.Electronic supplementary materialThe online version of this article (10.1186/s12943-018-0773-5) contains supplementary material, which is available to authorized users.

BackgroundA primary cutaneous melanoma will not kill the patient, but its metastases. Since in vitro studies on melanoma cells in 2-D cultures do often not reflect reality, 3-D models might come closer to the physiological situation in the patient during cancer initiation and progression.Methodology/Principal FindingsHere, we describe the chick embryo model for in vivo studies of melanoma cell migration and invasion. After transplantation of neural crest-derived melanoma cells into the neural tube, the melanoma cells resume neural crest cell migration along the medial and lateral pathways and finally undergo apoptosis in the target areas. Upon transplantation into ectopic areas such as the hindbrain or the optic cup malignant invasion and local tissue destruction occurs. In contrast, melanocytes are not able to spontaneously resume neural crest cell migration. However, malignant invasion can be induced in melanocytes by pre-treatment with the TGF-beta family members bone morphegenetic protein-2 or nodal. Transplantation of MCF7 breast cancer cells yields a different growth pattern in the rhombencephalon than melanoma cells.Conclusions/SignificanceThe chick embryo model is a feasible, cost-effective in vivo system to study invasion by cancer cells in an embryonic environment. It may be useful to study invasive behavior induced by embryonic oncogenes and for targeted manipulation of melanoma or breast cancer cells aiming at ablation of invasive properties.

Malignant melanoma has the highest propensity to metastasize to the brain of all primary neoplasms in adults. Here, we describe invasive growth and the development of melanoma metastases from suspensions of human melanoma cells in the brain of the chick embryo. Patient-derived melanoma cells and established melanoma cell lines were injected into the rhombencephalic brain vesicle of the two-day-chick embryo. After 48 and 96 h, tumor formation was studied in serial paraffin sections with melanoma-specific HMB45 and human-specific MIB1 proliferation markers. The majority of the cells injected into the embryonic liquor cavity perished. Only melanoma cells in newly formed aggregates or when attached to the dorsal roof plate escaped apoptosis. Local invasion occurred not in the ventral differentiating neural epithelium but only in the roof plate. Although after 48 h melanoma cells invaded the rhombencephalic roof plate profusely at both sides, after 96 h typically one large tumor developed in the midline between roof plate and the dorsal surface epithelium. From the tumor, single cells invaded the mesenchyme and blood vessels. Cell lines with different invasive properties retained their graded invasive behaviour. Maximally invasive cells formed continuous tracks via vessels and along nerve fibres. The central tumor in the roof plate of the chick embryo rhombencephalon resembles a metastatic nodule in the patients with melanoma. Penetration of the roof plate epithelium, tumor formation and invasion of surrounding tissues by single cells can reliably be reproduced. The chick embryo model can be used for molecular studies of early phases of melanoma brain metastasis.

No abstract

No abstract