Towards a melanoma phenoscape

Abstract

Owing to latest technical advances, tumor microenvironments can be resolved at single-cell resolution. Multiple molecular levels are measurable from individual cells in a dissociated as well as spatially-informative fashion. These powerful approaches are currently transforming our perception and understanding of cellular plasticity, a central feature of tumor biology that governs cancer progression and therapeutic resistance. Plasticity enables the development of dynamic “cell type flavors” (cell states). These cell states are defined as variations in molecular phenotypes within a cell type and anchored by characteristic transcriptional programs. The hunt for cancer cell states is on and has revealed already some shared programs across cancer types, reflecting to some extent the biology of their cell of origin. Melanoma, a neural crest-derived cancer, shows pronounced lineage identity with specific transcriptional programs reminiscent of melanocyte (de-)differentiation. Today, ecosystems of different cancer types are being charted on the single-cell level, often including also spatial information. In total, these atlas approaches profiled already hundreds of millions of cells across organs and diseases, paving the way towards next generation foundation models. We have contributed substantially to the knowledge of melanoma cell state heterogeneity and its TME composition by resolving melanoma samples under therapeutic pressure, as well as analyzing jointly human and murine melanoma samples in a cross-species approach. I will present an overview of our previous work, including an ongoing project that combines single-nuclei RNA-sequencing and spatial transcriptomics (Xenium) to map the primary melanoma architecture.