Spatial multiomics analyses of a cell atlas and interactome in skin cancer

Abstract

To identify the cells and oncogenic signaling pathways driving the initiation and progression of skin cancers, a knowledge base of not only cell type composition but also cell-to-cell interactions is essential. Using single-cell RNA sequencing, we generated a cell atlas of >50,000 cells from matched samples of healthy and squamous cell carcinoma (SCC) in the skin samples of 11 patients. We identified interactions that are enriched in cancer cells compared to cells in non-cancer samples. We further compared, at single-cell resolution, the cell-cell interactions between the three types of skin cancer (SCC, Basal Cell Carcinoma, and Melanoma) using CosMx Spatial Molecular Imager (SMI). Statistical integration of data across samples and replicates was performed to find differences in LR signaling between cancers, patients, and tissue regions. We found from two to 12 LR pairs that are specific to each of the three cancer types, some of which have been previously implicated in skin cancer progression and metastasis. Overall, we have identified 30 cell types and built an interactome with hundreds of thousands of possible interactions between these cell types at single-cell resolution. The unprecedented database suggests potential biological markers of initiation and progression of the most lethal type (melanoma) and the most common types (SCC and BCC) of skin cancer.