A Preclinical Model of Malignant Peripheral Nerve Sheath Tumor-like Melanoma Is Characterized by Infiltrating Mast Cells

2016 
Human melanomas show considerable variations in genetic changes, cell morphology and in microenvironmental composition. Genetically engineered mice have successfully been used to model the impact of genomic aberrations involved in melanoma pathogenesis. However, it is unclear whether they recapitulate the phenotypic heterogeneity of human melanoma cells and the complex interactions with the immune system. Here we report the unexpected finding that immune-cell poor pigmented and immune-cell rich amelanotic melanomas develop simultaneously in Cdk4R24C mutant mice upon melanocyte-specific conditional activation of oncogenic BrafV600E and a single application of the carcinogen DMBA. Interestingly, amelanotic melanomas showed morphological and molecular features of malignant peripheral nerve sheath tumors (MPNST). A bioinformatic cross-species comparison using a gene expression signature of MPNST-like mouse melanomas identified a subset of human melanomas with a similar histomorphology in the TCGA database. Exploring their transcriptional immune cell subtype compositions we found a highly significant association with mast cells. Importantly, mouse MPNST-like melanomas were also extensively infiltrated by mast cells and expressed mast cell chemoattractants similar to their human counterparts. A transplantable mouse MPNST-like melanoma cell line recapitulated mast cell recruitment in syngeneic mice demonstrating that this cell state can directly orchestrate histomorphology and microenvironmental composition. Our study emphasizes the importance of reciprocal, phenotype-dependent melanoma-immune cell interactions and argues for a critical role of mast cells in a subset of melanomas. We further conclude that our BrafV600E-Cdk4R24C model will facilitate the development of cell state-selective and microenvironment-directed therapies as it recapitulates at least two distinct human melanoma phenotypes at once.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    50
    References
    22
    Citations
    NaN
    KQI
    []